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

Abstract and Introduction


Background An association between eosinophilic esophagitis (EoE) and celiac disease (CD) has been suggested in the literature. Our aim was to confirm and quantify the association between these two diseases.

Methods All patients in a large Canadian city diagnosed with EoE or CD over a five-year period were identified. Standardized incidence ratios (SIRs) with 95% confidence intervals (CIs) were calculated.

Results Over the five-year study EoE was diagnosed in 421 patients and CD was diagnosed in 763 patients. The incidence of EoE ranged from 2.1 to 10.7 cases per 100,000 population. The incidence of CD ranged from 10.4 to 15.7 cases per 100,000 population. Among the EoE cohort, 83 (20%) cases of EoE and 245 (32%) cases of CD were diagnosed in pediatric patients. The incidence of EoE in the pediatric subpopulation ranged from 3.7 to 6.9 cases per 100,000 population. The incidence of CD in the pediatric subpopulation ranged from 9.5 to 22.7 cases per 100,000 population. The concomitant diagnosis of both EoE and CD was made in three patients, all of whom were pediatric males. The SIR for EoE in the CD cohort was 48.4 (95% CI = 9.73, 141.41) with a SIR for CD within the paediatric EoE cohort of 75.05 (95% CI = 15.08, 219.28).

Conclusions This study confirms the association between EoE and CD. However, this association may be limited to pediatrics where the risk of each condition is increased 50 to 75-fold in patients diagnosed with the alternative condition. The concomitant diagnosis of these conditions should be considered in pediatric patients with upper gastrointestinal symptoms.


Eosinophilic esophagitis (EoE) and celiac disease (CD) are distinct entities affecting the upper gastrointestinal tract, with different clinical and histopathological features. Although eosinophilic infiltration of the esophagus was first described in the late 1970s, it was not until 1993 that EoE was proposed as a distinct clinicopathologic syndrome.[1] Originally thought to be an uncommon condition, its incidence has dramatically risen over the past decade, recently estimated at 4.4 to 9.5 cases per 100,000 person-years.[2,3] When first described, EoE was felt to be a predominantly pediatric condition, however, it is now commonly diagnosed in adults as well as children.[4] There is a striking male predominance with 76% of adult and 66% of pediatric cases being diagnosed in males.[5]

The pathogenesis of EoE is not completely understood but appears to involve an interplay between genetic factors and environmental exposures. An association between EoE and allergic disorders such as asthma, atopic dermatitis, and food and environmental allergies has been long established.[12] In one pivotal study, a SNP in the gene that codes eotaxin-3, a cytokine critical to eosinophil migration, was associated with susceptibility to EoE.[11] Another GWAS has implicated over-expression of the gene that codes for thymic stromal lymphopoietin (TSLP) as an important factor in disease development. Interestingly, TSLP is a cytokine that regulates inflammatory response and has been implicated in other inflammatory conditions including asthma, inflammatory arthritis, and atopic dermatitis.[13–15]

Celiac disease is an autoimmune enteropathy that has been estimated to affect up to one percent of the general population, although many remain undiagnosed.[16] Recent incidence estimates range from 2 to 17.4 cases per 100,000 population.[17–20] The clinical features can be rather non-specific but malabsorptive features dominate in the form of diarrhea, iron-deficiency anemia and weight loss; in childhood there is failure to thrive and growth retardation.[21,22] While CD was also once considered a childhood disease, the majority of cases are now discovered in adulthood with an average age of diagnosis of 46 years.[21]

The development of CD is triggered by enteric exposure to the protein composite gluten in genetically predisposed individuals. The implicated genes at the HLA-DQ2 and/or DQ8 gene loci are found in 95 percent of people with celiac disease.[23–25] Recent genome wide studies have identified genes that convey overlapping susceptibility for a number of immune mediated conditions, including CD, type 1 diabetes, and Crohn's disease.[26] To date, there have been no SNPs identified that would suggest a common genetic basis for EoE and CD. Furthermore, the frequency of HLA alleles associated with CD are no more common in adults with EoE than the general population.[27]

The apparent absence of a genetic connection between EoE and CD is especially interesting given a clinical association that has been suggested in recent years. The first such reports were case studies of pediatric patients concomitantly diagnosed with both EoE and CD.[28–31] Two retrospective, cohort studies from Australia subsequently found a higher than expected prevalence of EoE among pediatric patients with CD. Leslie et al.[32] reviewed the medical records of 250 children with histologically confirmed CD over a seven-year period; 10 had been concurrently diagnosed with EoE. In a similar study, Ooi et al.[33] found that 7 of 221 children with CD had also been diagnosed with EoE. The estimated prevalence of EoE in these two pediatric CD cohorts was 4.0% and 3.2%, respectively. This is markedly higher than the 0.009% prevalence rate (0.9 per 10,000) reported for EoE in the general pediatric population.[34] Thompson et al.[35] recently reported on a cohort of 1142 adult and 297 pediatric CD patients presenting to a specialized CD referral centre in New York City and found an overall Standardized Incidence Ratio (SIR) for EoE of 16.0 using EoE incidence data from an Olmsted County population.

The incidence of autoimmune diseases of the gastrointestinal tract is increasing and despite excellent ongoing research the reason(s) remains unclear. As this is especially true of EoE and CD, we felt that it was of particular importance to confirm the association between EoE and CD using population-based pathology data. Given that the putative association has been mainly described in pediatric populations, we also sought to determine if this association extends to adults.