The most common endoscopic features in adults with EoE include linear or longitudinal furrows (80%), mucosal rings (64%), small caliber esophagus (28%), white plaques and/or exudates (16%), and strictures (12%) (23) (Fig. 40.1). In a large clinical series of 381 children,
the most common endoscopic features were linear furrows (41%), normal appearance (32%), esophageal rings (12%), and white plaques (15%) (18). It is important to note that the classic endoscopic features may be subtle and missed during endoscopy (16). Therefore, it is suggested that biopsies be taken for the clinical indication of unexplained dysphagia, refractory heartburn, or chest pain despite normal endoscopic findings (16).

While certain endoscopic features are characteristic of EoE, this condition is ultimately diagnosed by obtaining biopsy specimens which demonstrate histologic findings of increased intramucosal eosinophils in the esophagus without concomitant eosinophilic infiltration in the stomach or duodenum (16) (Figure 40.2). Other histologic features of this condition include superficial layering of the eosinophils, eosinophilic microabscesses (clusters of ≥4 eosinophils), intercellular edema, and degranulation of eosinophils. Other inflammatory cells such as lymphocytes, polymorphonuclear leukocytes, and mast cells may be present in the epithelium (29–31). Another histologic finding in EoE is epithelial hyperplasia, defined by papillary height elongation and basal zone proliferation (31). Epithelial hyperplasia is also a cardinal feature of the histopathology of reflux esophagitis. Studies have also shown presence of subepithelial fibrosis in biopsies of adults and children with eosinophilic esophagitis suggesting that deeper layers of the esophagus may be involved (3,32,34). Involvement of deeper layers of the esophagus has further been supported by the use of endoscopic ultrasound (35–37). It is speculated that this mucosal and submucosal fibrosis may lead to esophageal
remodeling and decreased compliance of the esophagus thus contributing to the symptoms of dysphagia even in the absence of an identifiable stricture.

Although a single diagnostic threshold of eosinophil density has not been determined, a recent consensus statement suggests using a threshold value of ≥15 eosinophils per high power field to diagnose EoE (16). It has also been demonstrated that the eosinophilic infiltration of the esophagus may not be evenly distributed within the esophagus (27). Therefore, it is suggested that biopsies be obtained from both the proximal and distal esophagus to obtain a higher diagnostic yield and perhaps increase the specificity of the diagnosis. A retrospective study of adult EoE patients found that obtaining more than five biopsies maximizes the sensitivity based on a diagnostic threshold of ≥15 eosinophils per high power field in the adult population (27). A follow-up study using a pediatric cohort demonstrated that 3 biopsies yielded a diagnosis of EoE in 97% of patients (38). In both the adult and pediatric studies, biopsies taken of only the proximal or distal esophagus missed the diagnosis in up to 20% of cases emphasizing the importance of taking biopsies from different locations.

Recent consensus recommendations based on a systematic review of the literature and expert opinion have led to the following diagnostic criteria. EoE is a clinico-pathological disease characterized by (a) the presence of symptoms including but not limited to dysphagia and food impaction in adults and feeding intolerance and GERD symptoms in children, (b) ≥15 eosinophils per high power field in the esophageal tissue, and (c) exclusion of other disorders associated with similar clinical, histologic, or endoscopic features such as GERD with either the use of high dose acid suppression prior to biopsy procurement or normal pH monitoring (16).

Although EoE is thought to be allergic, the allergen remains elusive. The concept of food as the allergenic cause was first introduced by Kelly et al. who described 10 children with severe GERD symptoms unresponsive to proton pump inhibitor (PPI) therapy (47). Eight patients had resolution and 2 had improvement in symptoms when placed on an elemental diet (ELED) for 6 weeks (Neocate One+, SHS North America, Gaithersburg, MD; EleCare, Ross Pediatrics, Abbott Laboratories, Abbott Park, IL). Histologically patients improved as well; eosinophil numbers decreased from a median of 41 per hpf (range 15 to 100) to a median of 0.5 (range 0 to 22) eos/hpf. The demonstration of recurrent symptoms on food reintroduction added credence to the role of food allergy. Markowitz et al. showed similar results in a study of 51 children diagnosed with EoE, using strict criteria to rule out GERD (48). For 3 months, 346 children with symptoms of GERD were treated with a PPI. Those who clinically responded to empiric PPI therapy or demonstrated an abnormal 24-hour pH study after treatment were excluded from the study. The remaining 51 children were treated with an ELED for 4 weeks with marked response. Eosinophil counts per hpf decreased on average from 34 to 1. Symptoms recurred with reintroduction of food. Similar results have been reproduced in several studies in children from multiple centers with larger cohorts showing resolution of symptoms and histologic evidence of EoE (18,49,51). However, elemental diets are unpalatable and very difficult to maintain and oftentimes require tube feedings. In the study by Liacouras et al., 80% of the children were fed by nasogastric tube and eight patients were unable to tolerate the diet (18). This poor tolerability has led to trials of empiric and allergy testing directed food elimination. Kagalwalla et al. retrospectively studied children with EoE, comparing those treated with an ELED to those treated with a diet eliminating the eight (six) most common food groups involved in allergy: nuts (tree nuts and peanuts), seafood (fish and shellfish), wheat, soy, milk, and eggs. This empiric approach was termed the six food elimination diet (SFED) (49). There were 60 children studied, 35 on SFED, and 25 on ELED. Twenty five of the 35 patients (69%) on SFED and 22 of 25 (88%) on ELED showed improvement in symptoms and resultant eosinophil counts ≤ 10 per hpf.

Work has been done to try to identify culprit foods for more directed rather than empiric food elimination. However, identification of the food with traditional methods of testing for food allergy has had limited success. Since EoE patients do not have reactions that are typical of immunoglobulin E (IgE)-mediated food allergic reactions such as urticaria or anaphylaxis, and since reactions may not be immediate, it has been difficult to identify responsible foods by history. In addition, studies show that symptoms do not necessarily correlate with disease (39). Children, however, tend to have more immediate symptoms such as vomiting and it is possible that their histories and allergy test results may be more reliable. Traditional tests used for determining food allergy, skin-prick test (SPT) and ELISA ImmunoCAP, are used to detect IgE antibodies and therefore may have a limited role in predicting food allergens in EoE. Furthermore, the sensitivity, specificity, as well as positive and negative predictive values for these tests, have shown substantial variability.