Eosinophilic oesophagitis and how to avoid them

EoE is clinicopathologic disorder and neither clinical nor pathologic information should be interpreted in isolation. Identification of dense eosinophilia in the squamous oesophageal epithelium is clearly an abnormal finding and the underlying cause should be identified;^1,2 however, oesophageal eosinophilia ≥15 eosinophils per high-power field (HPF) alone does not define EoE. Indeed, objective oesophageal eosinophilia in the absence of symptoms of oesophageal dysfunction (e.g. an incidental finding in patients with diarrhoea or in biopsy samples taken from patients with Barrett oesophagus) should be monitored, but a diagnosis of EoE should not be given without an adequate clinical context. In addition, several local and systemic diseases that have different clinical and histological features can be associated with oesophageal eosinophilia (e.g. eosinophilic gastroenteritis, achalasia, parasitic infection, hypereosinophilic syndrome, drug hypersensitivity, vasculitis, pemphigus, connective tissue disorders, graft versus host disease) and should be ruled out before a diagnosis of EoE is made.^1,2,6

EoE is a chronic inflammatory oesophageal disease that is triggered predominantly, but not exclusively, by food antigens. Therefore, it seems intuitive to perform food allergy testing to identify the triggering foods. Unfortunately, a testing-directed elimination diet had the lowest effectiveness rate in a meta-analysis of dietary interventions.¹⁰ These results were consistently low for studies performed in adults and variable among paediatric studies.¹⁰

Unlike IgE-mediated food allergy, EoE is a distinct form of food allergy that is largely driven by non-IgE delayed cell-mediated hypersensitivity.¹¹ Most skin and blood food allergy tests detect IgE-mediated responses. An atopy patch test can be used to elucidate delayed-type reactions to foods, but this test has not been standardized or validated for food allergy, including EoE. The accuracy of five different skin and blood food allergy tests to detect causative foods in adult EoE patients has lately been assessed.¹² None of the evaluated tests, measuring both IgE and non-IgE hypersensitivity, could accurately predict the causative foods previously identified in responders to an empiric six-food elimination diet (SFED).¹² Therefore, this diagnostic strategy should be discouraged in adults. For children, the highest efficacy (up to 53%) was reported in one single centre,¹³ but these results have not been replicated in other paediatric and adult studies.⁶

Untreated EoE is frequently associated with persistent oesophageal inflammation over time, leading to oesophageal remodelling that gives rise to stricture formation and functional abnormalities in the majority of patients. In a retrospective series of 200 Swiss adult EoE patients, the prevalence of fibrostricturing oesophageal features increased from 46.5% to 87.5% when the diagnostic delay increased from ≤2 years to >20 years.¹⁴ Similarly, diagnostic delay led to significant differences in oesophageal diameter in adult EoE patients, from <10mm with a mean delay of 14.8 years to ≥17mm with a delay <5 years.¹⁵ These results have been corroborated in a series from the US, in which the odds of having fibrostenotic features more than doubled for every 10-year increase in age.¹⁶

All these findings suggest that the natural history of untreated EoE is a continuum from an inflammatory to a fibrostenotic disease. Whether anti-inflammatory therapy (e.g. PPI, topical steroids or dietary therapy) can reverse the natural history of the disease remains to be elucidated. Recent studies have shown the ability of topical steroids and dietary treatment to reverse oesophageal fibrotic remodelling in children.^17–20

As it was explicitly included in the 2007⁵ and 2011¹ guidelines, many people still think that response to PPI therapy rules out EoE. GORD develops when the chronic reflux of stomach contents causes symptoms and/or complications, promoting a Th1 inflammatory response with recruitment of neutrophils and lymphocytes and a mild eosinophilic infiltration. The endoscopic appearance of the oesophagus may be normal in up to 80% of GORD patients. By contrast, EoE is a chronic immunoallergic disorder caused mainly by food allergens that promotes an aberrant Th2 inflammatory response, with eosinophil recruitment into the oesophageal mucosa. Typical endoscopic findings (e.g. rings, furrows, exudates, oedema and strictures) are present in up to 90% of EoE patients.

Evolving knowledge, mostly from adults, has demonstrated that patients with clinical and histological features of EoE that remit with PPI treatment (formerly called PPI-responsive oesophageal eosinophilia [PPI-REE]) are clinically, endoscopically, histologically, molecularly and genetically indistinguishable from EoE patients.²⁵ Aside from its anti-inflammatory effects, PPI monotherapy in PPI-REE patients also reverses the EoE abnormal gene expression signature, similar to the effects of topical steroids in patients with EoE. Some EoE patients who are responders to diet or topical steroids have also been shown to be responders to PPI therapy.^26,27 Accordingly, it seems counterintuitive to differentiate responders to PPI therapy from EoE patients based on a differential response to a drug (PPI therapy), when their phenotypic, molecular, mechanistic and therapeutic features cannot be reliably distinguished.

The recent description of EoE patients as responders to vonoprazan underscores the importance of acid reflux as a trigger of the disease.²⁸ Regardless of what drug patients are responsive to, responders to PPI therapy exhibit the clinical, endoscopic, histological, molecular and genetic features of EoE, (which are radically different from those of conventional GORD). These patients should not be labelled and treated as GORD patients, but rather as EoE patients.

Similar to that expected in inflammatory bowel disease, the ideal treatment endpoint for EoE would be complete resolution of clinical, endoscopic and histological features (deep remission) in order to prevent remodelling and related complications.^1,2,6 For this purpose, an induction phase, in which clinical and histological remission is achieved, should be followed by a maintenance phase, which is intended to prevent disease relapse and restore quality of life through sustained disease remission.³¹

A single therapeutic intervention should attempt to fulfil all the aforementioned therapeutic targets. Patients taking topical corticosteroids do not need dietary restrictions to be put in place, and topical steroid therapy should not be added for patients choosing dietary therapy. In addition to the potential unnecessary additive side effects and impairment of quality of life, an effective combination therapy may hinder getting to know which treatment was ultimately responsible for remission and which of the two treatments should be continued/discontinued for maintenance therapy. Likewise, combining different therapies in EoE studies might lead to results that are misleading and cannot be replicated.³² Evaluation of individual therapeutic interventions in EoE (e.g., PPI therapy⁹ or the SFED¹⁰) has produced consistent results in both children and adults. 

Treatment of EoE with an empiric elimination diet—the SFED—was first tested in Chicago in 2006.³³ This diet eliminated the six food groups most commonly associated with food allergy in the paediatric population in Chicago (cow´s milk protein, wheat, egg, soy, peanut/tree nuts, fish and seafood) for 6 weeks and led to clinical and histological remission in 74% of children.³³ Similar results have since been obtained in patients of all ages, as shown in a meta-analysis published in 2014.¹⁰ The effectiveness and wide reproducibility of the SFED are counteracted by the high level of dietary restriction and the large number of endoscopies required after reintroduction of individual foods. Less-restrictive empiric diets are therefore being evaluated.

Since three quarters of responders to the SFED have been found to have just one or two food triggers,³⁴ a four-food elimination diet (FFED), which avoids the most common food triggers (milk, wheat, egg and, to a lesser extent, soy/legumes) was developed. In the first prospective multicentre study in adult patients, the FFED achieved 54% remission,³⁵ whereas an abstract reporting the use of the FFED in a paediatric population revealed 71% efficacy.³⁶ Half of the responders to the FFED had one or two food triggers—cow´s milk and wheat were the most common.^35,36 Preliminary results have shown that a two-food elimination diet (cow’s milk and wheat) might achieve remission in 43% of children and adults, with one single food trigger identified in 70% of patients.³⁷

At present, most people still believe that the food groups included in empiric diets are removed from their regular diet indefinitely. In responders to any empiric 6-week diet, all food groups are reintroduced individually, with an endoscopy performed following each food challenge. The final goal is to provide a personalized maintenance therapy, with long-term removal solely of food triggers, namely, foods proven to induce oesophageal inflammation after individual reintroduction. 

Early findings for oesophageal dilation in EoE patients reported a high rate of complications, mainly oesophageal perforation and chest pain.^38,39 These findings were not confirmed in the first systematic review and meta-analysis of the literature, comprising 525 adult EoE patients and 992 endoscopic dilations.⁴⁰ Only three oesophageal perforations (0.3%) and one haemorrhage (0.1%) were reported, all at the same institution. Accordingly, the rate of major complications is consistent with that reported for endoscopic dilation in other oesophageal diseases (<1%).

Endoscopic dilation should be recommended to all EoE patients who have dysphagia/food impaction that is related to fibrostenotic abnormalities (either narrow-calibre oesophagus or strictures) and unresponsive to medical or dietary therapy.⁶ Endoscopic dilation is highly effective, with clinical improvement documented in 75% of patients in the aforementioned meta-analysis.⁴⁰ Mucosal lacerations after dilation should not be considered complications, but rather the intended outcome of the endoscopic procedure.