The Association Between Celiac Disease and Eosinophilic Esophagitis in Children and Adults

Michael J Stewart; Eldon Shaffer; Stephan J Urbanski; Paul L Beck; Martin A Storr


BMC Gastroenterol. 2013;13(96) 

In This Article


We undertook a population-based review of all patients diagnosed with CD and EoE within a large Canadian centre. The Calgary Health Region (CHR) is the sole provider of all medical and surgical services for 1.2 million people, 300,000 of whom are younger than 19 years.[36] A centralized laboratory and pathology service processes all adult and pediatric endoscopic biopsy specimens and generates histopathology reports that are stored in a searchable database. A keyword search was used to identify all duodenal and esophageal biopsies obtained within the CHR between January 1, 2004 and December 31, 2008. These pathology reports were then reviewed to determine if histological evidence of EoE or CD was present. For this review we defined the pediatric subpopulation as all persons younger than 19 years of age.

Eosinophilic esophagitis is a clinicopathological diagnosis based on the presence of typical symptoms and an esophageal biopsy demonstrating greater than 15 eosinophils per HPF.[5] Cases of EoE within our study population were identified by histopathology reports stating that the biopsy was consistent with or diagnostic of EoE.

The diagnosis of CD was based on duodenal biopsies demonstrating typical histopathological features and classified using the modified Marsh criteria.[37] In general, biopsies demonstrating Marsh stage 2, 3a-c, and 4 lesions are diagnostic of CD. Marsh stage 1 biopsies are less specific and only considered to represent CD in the setting of a positive serological marker. (e.g. elevated anti-tissue transglutaminase). At the time of this study there where no formal requirements for pathology reporting of duodenal specimens, therefore, pathology reports that did not specify a Marsh stage were considered positive for CD when the pathologist commented that the biopsies were consistent with, or diagnostic of CD. Serologic confirmation of CD was obtained if pathologist comments were indeterminate (e.g. Marsh stage I lesions) and determined to represent a positive case of CD is a serologic marker was consistent with the diagnosis of CD and negative if the serology was normal or unavailable.

Statistical Analysis

Crude incidence rates were calculated using CHR population data for the years 2004–2008.[36] Crude incidence rates were age- and gender-adjusted using the age and gender distribution of the 2006 Canadian population[38] as the reference population. This adjustment was performed in order to mitigate the effect of any regional variability in age and gender distribution.

To quantify a relationship between EoE and CD, SIRs with 95% confidence intervals (CIs) were calculated.[39] A SIR is an estimate of the true occurrence of disease in a population relative to what might be expected based on incidence rates in an average ("general") population. The ratio comes from dividing the observed number diagnosed (the number of cases of a specific disease) within a reference population by the expected number within the general population. A SIR greater than unity (1.0) would indicate a higher than expected number of cases within the reference population as compared to the "general" population.

This study was approved by the Conjoint Health Research Ethics Board at the University of Calgary and is in agreement with the Declaration of Helsinki.