Eosinophilic esophagitis (EoE) is a chronic immune-mediated disease with an incidence rising over time, estimated at 1/2500/year, which affects children or young adults.1 Clinical presentation, related to esophageal dysfunction, varies with age: non-specific symptoms (feeding difficulties, vomiting, failure to thrive…) in children under five years old, whereas older children often relate dysphagia or food impaction. Histologically, pathology found an esophageal infiltration by eosinophils, with a cluster at 15 eosinophils/high-power-field (HPF).2 EoE is often associated with atopic diseases and food allergies seems to be involved.3 Thus, since 2007, guidelines recommended allergy testing of these patients.2 However, the utility of allergy testing is still debated and not yet codified. Moreover, invasiveness of endoscopy limits management in children which is not well determined, emphasizing the need for harmonization. To optimize our practices, we analyzed our pediatric population and its management, with a focus on allergy assessment, in comparison with the literature.

This is a retrospective review of charts of children, from 0 to 18 years of age, with a diagnosis of EoE from January 1, 2006 to January 31, 2016 in our pediatric center (Hôpital Femme Mère Enfant, Bron, France). EoE was defined as ≥15 eosinophils/HPF (High-Power Field) in at least one esophageal biopsy. The study was approved by the local review board. A p-value <0.05 was considered as statistically significant.

Demographically, our population was consistent with the literature with a predominance of males (79.6% in our cohort versus 70–85% in literature) and young children (median age was nine years versus 5–10 years in literature)4,5 with a frequent atopic background (70.3%). As reported in previous works, familial atopic background was reported in 46.3% of our patients. Genetic variation was analyzed in literature, and some clusters were recently identified.6 Other associations have been reported (esophageal atresia, transplantation, EBV infection, preterm babies…) but the results were not significant in our cohort, perhaps due to the small sample size.

First of all, as observed in literature, symptoms in our cohort differed according to the age: child under five years presented with reflux symptoms, vomiting, dysphagia, feeding difficulty and failure to thrive, whereas older children and adolescents more likely presented with food impaction, abdominal pain and dysphagia.7,8 These two patterns of symptoms could be linked to the natural course, with a first inflammatory phase followed by fibrosis.9 Cough is rarely reported in the literature but appeared as a frequent symptom in our cohort, especially between three and 10 years and should be considered as a suggestive symptom of EoE in young children.

The delay in diagnosis (median of 6.65 years in our cohort) was consistent with literature (six years10) and can be explained by the non-specificity of symptoms, an underestimation by the patient himself, the invasive character of explorations and the overlap with gastroesophageal reflux disease. Moreover, this diagnosis could be miss diagnosed due to a normal aspect in endoscopy (20% and 7% of patients in retrospective prospective analysis, respectively11) and focal infiltration requiring multiple biopsies.12 Thus, the diagnostic sensitivity of a single biopsy was 73% in children and increased above 97% after five biopsies, so guidelines recommend at least six biopsies, including the three levels of esophagus.2 The majority of our patients benefited from these multiple biopsies, whereas endoscopy was normal in 12.9%. An endoscopic score, called EREFS, was proposed to classify and grade EoE, with great specificity (90–95%) but with low sensitivity and negative predictive value of endoscopic features (20–30%).13

The cut-off of 15 eosinophils/HPF was arbitrarily determined in 2007 guidelines, and then demonstrated 100% sensitivity and 96% specificity.14 Other signs in pathology included micro-abscesses, basal zone hyperplasia, dilated intercellular spaces, eosinophil surface layering, papillary elongation, and lamina propria fibrosis. A large majority (95.4%) of our patients presented with these other pathological signs.

Currently, diagnosis and follow up of disease activity relies on endoscopy, involving anesthesia (with a potential risk of general anesthesia on cerebral development15) and invasive methods. This can be an obstacle in the management of young children. However, none of the other exams (esophagram, manometry, EndoFLIP, biomarkers…) were actually valuable.

Clinical scores could be interesting to predict and follow EoE but dissociation between symptoms and endoscopic/histologic features limits their utilization.16 In our cohort, 13% of the patients were asymptomatic but with positive pathology.

In our center, allergy testing was a central element to guide the management of patients, including atopy patch tests, skin prick tests and IgE levels. Our frequent food allergens are concordant with literature, adding chicken, with prick tests positivity in 27.3% and IgE in 60.0% in our cohort.

Skin prick tests demonstrated a low positive predictive value (average of 47%) but a really better negative one (>90%).17 Thus, the principal goal of skin prick tests is not to determine the culprit antigen but to determine non-implicated ones to limit diet exclusions and their impact on growth and deficiencies. This is particularly important in EoE because major allergens (milk, wheat, egg, fish, nuts, peanuts and soy…) represent an important part of a child’s diet.

The use and interpretation of atopy patch tests are not standardized in EoE because concordance between patch test results and EoE’s food triggers identified by biopsy-monitored sequential food reintroduction is low, with <50% sensitivity, except for milk (86%). However, they are frequently used in pediatric cohorts where results are better than in adults with a positive predictive >90%. Thus, in our experience, atopy patch tests for milk, wheat, potatoes, egg, beef, and chicken proved their benefit and have the advantage of testing non-immediate allergic reactions (physiopathology of EoE staying unclear with potential limited relevance of IgE). Moreover, the combination of prick and atopy patch tests increases the sensitivity rate to 65–95% and Henderson et al. found that combination to be effective to guide therapy.18 Thus, we tend to generalize the use of atopy patch tests.

As for the prick test, the negative predictive value of IgE testing in better than the positive one but concordance with symptoms is low.19 Several studies showed that patients with low or undetectable IgE could respond to the eviction of incriminated food. Thus, recombinant humanized anti-IgE antibody (Omalizumab) seems ineffective to treat EoE.20 Notwithstanding, IgE can be useful to identify culprit allergens in case of a significant increase.

Chip was evaluated to guide diet and effective in 66.7% of our patients but the literature results are not convincing, with 93% of patients who failed to achieve endoscopic remission.21 However, these results concern an adult population and results of allergy testing are different between an adult and a pediatric cohort.

Other tests (MAST CLA, serum IgE levels, blood eosinophils count…) described in our cohort were not recommended in guideline because of non-encouraging results. We found no relation between blood eosinophilic count and esophageal eosinophilic count.

Therefore, allergy testing cannot affirm the diagnosis but could sometimes help to identify foods and undoubtedly guide dietary therapy to resolve symptoms and avoid unnecessary exclusions (to limit the impact on growth and deficiencies but also limit the risk of subsequent anaphylaxis in case of prolongated exclusion in sensitized allergens).

The dietary approach is sustained by the efficacy of amino acid-based formula in resolving the disease with a remission rate of 90.8% but its implementation is difficult because of poor taste, social impact, cost… Therefore, some authors suggest the use of elemental diets only after failure of other diets. This diet was used in only 4.7% of our cohort. Empiric diets eliminate common food allergens without prior allergy testing. Most common are 1, 2 or 6-Food Elimination Diets, including most common allergens (milk, wheat, egg, soy, peanut and nuts, and shellfish and fish), with histological remission in 65–77%.22,23 Most children’s EoE is triggered by one to three foods, so better management would be a sequential reintroduction and monitoring of symptoms and biopsy after each change, but the necessity of repeated endoscopy limits their utilization (thus, none of our patient followed these diets).

As demonstrated by Spergel, the dietary restriction based on prick and patch tests seems to be useful, and has the advantage of excluding fewer foods.24 However, it demonstrates some limits, especially a poor predictive positive value, with clinical and histological response in only 48% of child.25 This type of diet seems to be especially effective in the mono-sensitized patient.

The goal of treatment is the resolution of inflammation to prevent disease relapse and long-term complications such as fibrosis or strictures. Several forms of therapy are currently used in practice, in addition to dietary eviction or no.26

Medical treatment includes PPIs (representing now the first line of treatment, with histologic remission and symptom improvement in 50 and 60%, respectively (71.5% in our cohort))2,27,28 and topical steroid (with equivalent efficacy and fewer adverse events than systemic corticosteroid). Fluticasone propionate swallowing, and recently viscous budesonide induced histological remission in 50–90% of patients.2,29 However, the clinical response is less clear and two recent meta-analyses did not find superiority to placebo.30

Other treatments, such as immunosuppressive treatment, Omalizumab… were not used in practice.

In our study, only 2.8% of patient needed esophageal dilation. Risks are poor if dilatation is progressive but if a clinical improvement at one year is noticed in 85% of the patients, the benefit is not effective in the long-term.

A favorable evolution was found in 72.7% of our patients, especially among mono-sensitized patients and the allergy-testing group.

Due to the lack of specific recommendations in children, especially regarding allergy testing, EoE management remains very heterogeneous. Management of these children has to be standardized in pediatric centers according to adult guidelines and local practices. This report underlines some interesting facts in our population such as the frequency of chronic cough in children <10 years, advantage of allergy testing (including meat tests) with a role of atopy patch test and the need for harmonization of testing and treatment.

The authors have no conflict of interest to declare.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.