HOME
Return to Table of Contents
DEVELOPMENT OF THE CELIAC DISEASE ACITIVITY RATING SCORE (CeDARS)
A Rostom1, J Bai, C Catassi, A Fasano; the CeDARS investigators 1University of Calgary
Background and Aims: Celiac Disease (CeD) is a common and underdiagnosed condition of intestinal inflammation related to gluten ingestion in genetically susceptible individuals. While significant progress has been made for the accurate diagnosis of this condition, the monitoring of disease activity remains an important challenge. While several candidate measures exist such as histology, or serology, each has significant shortcomings for monitoring activity. The purpose of this study was to develop a composite activity rating score that can be used to monitor and rate CeD activity.
METHODS: A list of candidate variables was circulated electronically for comments and initial voting. Experts were asked to consider these on the basis of: 1) presumed association with CeD activity, 2) ability to reflect activity over a spectrum of disease presentations and duration; 3) reliability and responsiveness. The group was assembled in November 2006 in New York City. A modified Delphi technique was used to review the candidate variables, and to review and update the voting. The experts were asked to vote on whether, a given variable should be included, and to rank their top 6 variables. A candidate variable was included in the short list, if greater than 50% of the group voted to include it. Several data sets were obtained from previous studies to retrospectively develop the CeDARs. Univariate analysis appropriate to the variable type was used to identify components of disease activity that were statistically significantly different in celiac patients in “remission” versus those with “active disease”. Stepwise multivariate logistic regression was then used to identify those variables identified in the univariate analysis that independently predicted disease activity. The model that produced both a high sensitivity and specificity of (>90% for classifying patients in to the remission or active disease groups), and appeared to best explain the data variability was chosen as the candidate model.
RESULTS: 204 observations were available that included 125 remission and 79 active disease observations. On univariate analsysis, histology (multiple measures), intestinal permeability, serology (AGA, tTG, EMA), BMI, skin fold thickness, SF36, and GSRS statistically distinguished activity from remission. On multivariate logistic regression; several potential models arose with sensitivity and specificities greater than 90% for further prospective validation.
CONCLUSION: CeDARS has been development and is currently undergoing prospective validation alongside 2 RCTs.