The relationship between eosinophilic esophagitis and immunotherapy

Introduction

Eosinophilic esophagitis (EoE) is a chronic disease associated with type 2 inflammation. EoE usually requires long-term surveillance with esophagogastroduodenoscopy (EGD) and treatment with medications or elimination diets. In severe or untreated cases, fibrostenosis and esophageal stricture can occur.¹ The causes of EoE are unknown, however, recent studies suggest it may occur in patients receiving treatment with immunotherapy.^2–5

Immunotherapy has been employed for over a century as a treatment for allergies.^6,7 While the antigen, route, timing, and mode of delivery may differ, immunotherapy is based on the principle that gradual, incremental allergen exposure induces desensitization. Interestingly, some forms of immunotherapy are being employed as a treatment for EoE,^3,8–11 while others may cause or exacerbate it.^2–5 Oral immunotherapy (OIT) specifically induces esophageal eosinophilia which may be transient or persistent. Once symptoms of esophageal dysfunction occur in association with esophageal eosinophilia, the disease may be reversible with OIT discontinuation or become permanent. The development of esophageal eosinophilia and EoE in the setting of controlled exposure to a trigger antigen may provide unique insights into EoE pathogenesis as esophageal eosinophilia or reversible disease may be precursors to the chronic disease state. The purpose of this review is to examine the relationship between EoE and immunotherapy. We will primarily focus on EoE and OIT, as this association is most prevalent in the literature, but we will also briefly discuss other forms of immunotherapy.

Definitions of Immunotherapy

There are multiple forms of immunotherapy, including oral (OIT), subcutaneous (SCIT), sublingual (SLIT), and epicutaneous (EPIT) immunotherapy. OIT is an elective treatment taken daily to reduce the risk of anaphylaxis in IgE-mediated food allergy. SCIT is indicated for patients with allergic rhinoconjunctivitis and/or allergic asthma and is administered monthly after a maintenance dose is reached, requiring 3–5 years of injections to maintain aeroallergen desensitization. SLIT is another method for aeroallergen desensitization that can be administered as a tablet or liquid extract. SLIT, administered as sublingual drops, and EPIT, applied as a patch, are being studied for food allergen desensitization.

Background

Epidemiology of OIT-Induced EoE

The estimated incidence and prevalence of EoE during OIT range from 1–3.2% and 2.7–6.9%, respectively^2,12–14 (Table 1). Lucendo et al initially summarized development of EoE in a systematic review reporting a 2.7% incidence amongst those receiving milk, peanut, and egg OIT over a nine-year period. Twenty patients had EoE confirmed by esophageal biopsy (milk OIT = 10, peanut OIT = 7, egg OIT = 2), while 1 patient developed EoE with multiple allergen OIT.² Morales-Cabeza et al analyzed data over a 15-year period in a pediatric Spanish cohort receiving milk, egg, or peanut OIT and reported a prevalence of 2.8%. Of these, 10 and 6 patients were receiving milk and egg OIT, respectively, while 1 patient received both milk and egg OIT.¹² Petroni et al assessed symptoms reported during OIT and found that the incidence of biopsy-confirmed EoE was 5.3%. Of these, 4.2% of cases were induced by egg, 5.4% by milk, and 5.2% by peanut OIT.¹³ Longitudinal studies focusing on peanut¹⁴ and milk OIT¹⁵ reported an overall incidence of 1% in a cohort of 1,217 participants and 6.9% in a cohort of 1,545 pooled participants.¹⁴ EoE typically developed in children under 18,^2,12–14 with one reported case in a 33-year-old female; however, there are limited studies of OIT in adults.¹⁴

Table 1:

Incidence and prevalence of EoE for patients receiving OIT

Author	Incidence	Prevalence	Total number of patients reported to have EoE
Lucendo et al. (1)	2.72%	2.7%	20 out of 711
Morales-Cabeza et al. (2)	-	2.8%	17 out of 607
Petroni et al. (3)	-	5.6%	35 out of 663, biopsy confirmed
Nilsson et al. (4)	1%	-	12 out of 1217
Bognanni et al. (6)	-	6.9%	Number of patients with EoE unknown, total patients 1545
Dunlop et al. (7)	3.2%	-	6 out of 187

The most common symptoms reported included abdominal pain and vomiting,^2,12–14 prompting endoscopy and diagnosis of EoE. The median time to symptom development was 25 months after the up-dosing/build-up phase, with EoE histologically confirmed by 36 months. Although esophageal biopsies were not reported across studies, eosinophilic abscesses, basal cell hyperplasia and lamina propria fibrosis were noted.¹²

The natural progression of EoE in patients receiving OIT is unclear. Studies are limited by lack of esophageal biopsies prior to OIT or before symptom manifestation.^2,12–14 As a result, discerning whether eosinophilia existed before OIT initiation or whether OIT exacerbates EoE is challenging. This is due to lack of recognition by patients and symptom screening by providers, along with use of proton pump inhibitors (PPIs) for abdominal discomfort masking symptoms in earlier studies.^2,13 Additionally, whether esophageal barrier dysfunction is present or develops in patients on OIT is unknown. Studies on serial EGDs of patients receiving peanut OIT revealed esophageal barrier dysfunction and eosinophilia. Despite this, not all patients developed EoE and eosinophilia resolved in some.¹⁴ Thus, major questions remain regarding the contribution of OIT towards the development of EoE.

Multiple other factors are associated with EoE development in this subset of patients. Allergic predisposition (e.g. asthma, allergic rhinitis, or other food allergies), family history of EoE, and male gender were prevalent in these patients.^2,12–14,16 Importantly, as the disease did not always improve with removal of the allergen from the diet,^12,14,16 intrinsic cellular or genetic factors may also contribute.

Lastly, milk, egg and peanut OIT are more likely to precipitate the development of EoE, with milk being most common food trigger. In a study of children receiving milk OIT, 6.9% of patients developed EoE.¹⁵ Likewise, the introduction of baked milk after food challenge in patients with IgE-mediated milk allergy led to the development of EoE in 3.2% of patients.¹⁷ Whether this reflects a greater proportion of patients on milk OIT, compared to other allergen OIT, versus an inherent ability of milk allergens to skew the esophageal inflammatory response towards EoE is unknown. Single-cell studies of TCR diversity show clonal expansion of pathogenic effector Th2 cells in the esophagus of EoE patients. A subset of these pathogenic effector T cells in the peripheral blood were reactive to milk antigen, suggesting that the immunogenicity of milk allergens may predispose to EoE.¹⁸ More studies that examine the molecular basis of allergen sensitization in EoE are needed.

Discussion

Potential Explanations for EoE During OIT

Pre-existing Disease

There are several possible explanations for the occurrence of EoE during OIT. First, a patient may have undiagnosed, pre-existing EoE at OIT initiation. Inadequate screening for gastrointestinal symptoms before starting OIT may allow EoE to go undetected.¹⁹ Indeed, baseline gastrointestinal symptoms are relatively common among atopic individuals, and previously undiagnosed EoE has been diagnosed in these patients. Among 100 adults with atopic disease but without known EoE, 44% reported solid food dysphagia ≥ 1 time/week and/or food impaction for ≥ 30 minutes in the preceding year.¹⁹ Additionally, within the preceding week, participants frequently reported other gastrointestinal symptoms (abdominal pain 42%; nausea/vomiting 19%; reflux 43%). Cytosponge^™ was performed in 5/44 patients with dysphagia, and EoE was detected in one. One patient with negative Cytosponge^™ findings was later diagnosed with EoE on EGD. Two participants with other gastrointestinal symptoms were diagnosed with EoE by EGD. In another study of 89 children and adults with history of cow’s milk-related anaphylaxis, 37 (42%) reported gastrointestinal symptoms, and 34 (38.2%) had > 15 eosinophils/high power field (eos/hpf).²⁰ Patients with significant esophageal eosinophilia were treated with a PPI for 8 weeks, then underwent a second EGD. After the 8 week PPI trial, ten still had > 15 eosinophils/hpf, with 9/10 reporting gastrointestinal symptoms.

Some patients may have eosinophilic inflammation in the absence of symptoms of esophageal dysfunction. Asymptomatic esophageal eosinophilia can be an incidental finding when EGD biopsies are performed for another indication. Baseline esophageal eosinophilia (>15 eos/hpf) was noted in 14% (3/21) of adult subjects with IgE-mediated peanut allergy, prior to peanut OIT initiation in a small, prospective, longitudinal cohort.²¹ None reported dysphagia, and some reported mild gastrointestinal symptoms (abdominal pain in 24%, mild reflux not associated with food in 5%), but none of those with symptoms had esophageal eosinophilia. It is possible that patients with asymptomatic esophageal eosinophilia are at higher risk for EoE development during OIT; however, this has not yet been established.

EoE Develops Independent of OIT

Some patients may develop EoE regardless of immunotherapy. EoE prevalence is higher in food allergic patients, and underlying atopic conditions increase the risk of developing EoE.^22,23 Having ≥ 1 underlying atopic condition is associated with EoE diagnosis.²³ The prevalence of EoE in children with IgE-mediated food allergy is 4.7%,²² and estimates of OIT-induced EoE are similar (5.3%).¹³ While these prevalence rates are similar, the temporal relationship between OIT initiation and EoE onset should be considered. Of note, EoE prevalence may be higher in patients receiving OIT due to detection bias, as OIT typically requires clinic visits every 2 weeks and providers maintain a high index of suspicion due to the known association between EoE and OIT.

OIT Induces EoE Development

The final and most probable explanation is that OIT can induce EoE. Specific symptom and laboratory criteria have been proposed to monitor gastrointestinal manifestations during OIT in the absence of biopsies. Goldberg et al described a clinical entity, Oral Immunotherapy-Induced Gastrointestinal symptoms and Eosinophilic Responses (OITIGER), defined as abdominal pain or vomiting ≥ 3 days/month, not related to OIT dose timing, with absolute eosinophil count (AEC) ≥ 900 eosinophils/μL.^24,25 In their initial study, over 8% of OIT patients met symptom criteria for OITGER.²⁴ A higher maximum AEC was associated with an increased likelihood of gastrointestinal symptoms. When compared to asymptomatic subjects, those with OITIGER had a higher baseline AEC. All 3 patients who underwent EGD met histologic EoE criteria.

In a subsequent study, all patients with OITIGER had symptomatic resolution and decreased AEC after cessation or reduction of OIT dosing.²⁵ Seventeen of 65 patients with OITIGER permanently discontinued OIT. When compared to asymptomatic patients, the OITIGER group were less likely to reach full food allergen desensitization, although over 70% achieved doses allowing partial desensitization. After restarting OIT, more than 18% had return of OITIGER. OITIGER was more common in patients undergoing milk vs. peanut OIT. In milk OIT patients, OITIGER was more likely with a higher initiation dose (> 120 mg), increased dose escalation (> 4-fold increase above the starting dose), and AEC > 600 eosinophils/μL. It is important to note that in both studies by Goldberg et al, the majority of patients with OITIGER did not undergo EGD, likely underestimating EoE incidence.

Wright et al reported a randomized, double-blind, placebo-controlled study demonstrating a likely causal relationship between EoE and OIT.²⁶ EGD was performed in subjects receiving peanut OIT vs. placebo at baseline, 52 weeks, and 104 weeks. After 52 weeks, esophageal eosinophilia (≥ 15 eos/hpf) was induced or increased in 57% (4/7) of OIT patients. One subject was formally diagnosed with EoE and withdrew after developing a food impaction. Interestingly, esophageal eosinophilia was usually transient and was dissociated from clinical symptoms. None of the patients treated with placebo met histologic criteria for EoE. In the same cohort, dilated intercellular spaces were noted in all patients at baseline, suggesting patients with IgE-mediated peanut allergy may have esophageal barrier impairment before initiation of OIT.

Outcomes after Discontinuation of OIT

Response to discontinuation of OIT after the development of EoE is variable. For a subset of patients, removal of the inciting food allergen was sufficient to resolve EoE.² In other patients, stopping OIT alone did not lead to remission, and patients required PPIs and/or oral steroids to achieve remission.^12,14,16 For patients choosing to continue OIT, incorporation of PPIs and oral steroids led to remission of disease, which continued through median follow-up at 50 months.¹² In fact, among 17 patients who developed EoE while on milk and egg OIT, 4 patients discontinued OIT without clinical response and required a PPI, while 11 patients continued OIT with a PPI. Only 1 patient required oral steroids prior to remission.¹² In a separate cohort of patients on peanut OIT, discontinuation of peanut was only implemented for 1/12 patients with 10 eventually requiring PPIs and/or oral steroids.¹⁴ In this study, 5 patients achieved remission, while 2 continued to have active EoE on repeat EGD.¹⁴ Predictors of persistent OIT-induced EoE have not been identified.

Aeroallergens and EoE

Animal models first suggested a potential link between aeroallergens and EoE in 2001, when significant esophageal eosinophilia was observed following intranasal exposure of Aspergillus fumigatus in mice, resembling the histologic pattern of EoE.^5,27–29 Subsequently, case reports demonstrated a seasonal correlation with EoE symptom exacerbation. One described a 21-year-old female with asthma and allergic rhinoconjunctivitis, who developed worsening EoE symptoms and esophageal symptoms that correlated temporally with pollen exposure.²⁸ Additionally, Onbasie et al demonstrated esophageal eosinophilic infiltration in 26% of grass pollen-sensitized patients with allergic rhinitis during the pollen season.³⁰

Further substantiating this relationship, a cohort study of 127 adults diagnosed with EoE between January 2006 and November 2008 unveiled a distinct seasonal pattern for newly diagnosed disease. The highest incidence of EoE diagnoses was recorded during the spring (April-June), accounting for 33% of cases (binomial p=0.023), while winter months (January-March) exhibited a marked reduction at 16% (binomial p=0.01). Diagnoses made during the summer (July-September) and fall (October-December) demonstrated no significant differences, comprising 26% and 24% of cases, respectively.

Ram et al also identified a seasonal trend in EoE exacerbations. In a retrospective review of 1,180 pediatric patients with EoE, 160 (14%) exhibited an exacerbation of symptoms triggered by aeroallergens with histologic confirmation of esophageal eosinophilia during pollen seasons in 32 of these patients (20%).³¹ Moreover, a positive correlation was identified between the peak grass pollen count and EoE cases during the spring (r_s = 1.0, p<0.01). The precise nature of aeroallergen involvement in EoE pathogenesis is uncertain. It is unclear whether individuals with pre-existing EoE experience worsened symptoms during periods of elevated outdoor pollen concentrations or if these instances represent entirely new EoE diagnoses.²⁷

Aeroallergen SCIT and EoE

Published case reports have advocated for the consideration of SCIT as a potential therapy for a specific subset of patients with EoE and allergic rhinoconjunctivitis or oral allergy syndrome.^10,32 In some of these cases, SCIT initially exacerbated symptoms and eosinophilic inflammation; however, continuation led to gradual symptomatic improvement and possible disease remission over time.³² Ramirez et al also described a 4-year-old boy with refractory EoE and allergic rhinoconjunctivitis who underwent dust mite SCIT followed by resolution of persistent emesis over the ensuing two years. EoE remission was confirmed by biopsy.¹⁰ In adults, Perez et al described a 65-year-old man with EoE and seasonal allergies who demonstrated resolution of EoE symptoms after 5 years of SCIT intended for uncontrolled environmental allergies.^8,31,32

Robey et al examined the use of SCIT in the setting of EoE in a retrospective cohort study. Ten adult patients with EoE who underwent SCIT (EoE+SCIT) were compared to 667 patients with EoE who did not undergo SCIT (EoE-SCIT). Both groups had similar baseline characteristics and endoscopic features, but the EoE+SCIT group exhibited longer duration of symptoms before EoE diagnosis (13.8 vs 7.3 years, p=0.046) and had more atopic disease.¹¹ Initial EoE treatment with swallowed steroids in the EoE+SCIT group resulted in a symptomatic response in 60% and histologic response in 30% of the patients. After the introduction of SCIT, of the 5 patients who had follow up data and were concurrently treated with SCIT, 40% had symptomatic response and 40% had histologic response. There were no adverse events reported in the EoE+SCIT cohort. This study suggests that SCIT in patients with EoE is safe and may be worth considering in some patients.¹¹ Together, this literature suggests a link between aeroallergens and EoE, with the potential consideration for SCIT as an avenue for select patients; however, controlled studies are needed to validate these findings.

Aeroallergen SLIT-Induced EoE

Significant advancements have been achieved in the development of therapeutic strategies for atopic conditions, notably SLIT. However, advancements are occasionally accompanied with unanticipated outcomes, and a recognized link between EoE and aeroallergen SLIT has raised clinical concerns. As a result, EoE is a contraindication to the administration of SLIT therapy.^4,29

The first report of eosinophilic gastrointestinal disease (EGID) following SLIT was in 2013, involving a 44-year-old woman with new-onset dysphagia 4 weeks after starting SLIT for hazelnut, birch, and alder. EoE was confirmed on biopsy. She experienced resolution of symptoms and esophageal eosinophilia within 4 weeks of SLIT discontinuation with sustained symptom resolution by 12 weeks. This case stands as a significant milestone in highlighting the potential interplay between SLIT and the development of EoE.^3,4

Since then, additional reports have emerged detailing a similar connection. Cafone et al and Votto et al summarized case reports of SLIT and EoE development. These reports included patients ages 9 to 53 years and observed that the onset of EoE occurred during both the up-dosing and maintenance phases of SLIT. After histologic confirmation of EoE, subsequent remission occurred 1 to 16 months following discontinuation of SLIT.^3,4 These reports emphasize the clinical relevance of EoE and SLIT, underscoring the importance of individualized consideration prior to treatment with SLIT. In cases where gastrointestinal symptoms manifest during SLIT treatment, timely assessment is imperative to ensure prompt diagnosis, appropriate treatment, and discontinuation of potentially problematic therapy.³

Proposed Options for Management

General Considerations

The generally accepted clinical practice is to discontinue OIT or SLIT in patients newly diagnosed with EoE. However, recent publications report simultaneous OIT continuation with EoE treatment. ^33–35 In a pediatric OIT cohort, 13/1994 (0.7%) patients developed EoE. ³⁴ Of these, 12 (92%) continued OIT. One discontinued OIT due to OIT-associated anaphylaxis and had spontaneous EoE resolution. Another did not receive EoE treatment, continued OIT, and had symptomatic and histologic EoE resolution on follow-up EGD. Eleven patients were treated with high-dose PPI monotherapy, and 5 achieved EoE remission. Of the remaining 6 patients, 2 received swallowed topical corticosteroids, 2 continued low-dose PPI, and 2 discontinued OIT. One patient receiving swallowed topical corticosteroids had complete symptomatic and histologic remission. The other patients had clinical resolution and only partial histologic resolution.

As gastrointestinal symptoms during OIT are relatively common and may or may not indicate underlying EoE/EGID, Chua et al developed an algorithm to manage and risk-stratify these patients.³⁵ If symptoms are moderate/severe or if EGD is readily available, and the caregiver is in agreement, OIT dosing should be paused until gastroenterology evaluation. If this is not the case, a PPI trial may be initiated and OIT dosing may be paused vs. continued at the same or reduced dose. If effective, PPIs can be discontinued, and OIT dosing is increased more gradually. Pending further research evaluating long-term outcomes of OIT continuation after EoE diagnosis, the decision to deviate from current standards of care should be carefully considered after shared decision-making between the patient/caregiver and provider.³⁶

Milk EPIT for Milk-Induced EoE

EPIT has been used to treat IgE-mediated food allergy through delivery of antigen through a transdermal patch. Cutaneous reactions are the most common side effect and EPIT-induced EoE has not been reported. In fact, milk EPIT has been proposed for management of milk-induced EoE. Following evidence of successful reduction of EGID with EPIT in mice and pig models, Spergel et al performed a randomized placebo-controlled trial to study the efficacy of Viaskin milk-EPIT (500 mg of milk protein) in 20 children with milk-induced EoE. Subjects were randomized in 3:1 ratio (Viaskin Milk to placebo) and were on milk-free diet for 9 months, followed by milk-containing diet for 2 months. Histologic endpoints demonstrated no significant difference in mean eos/hpf in the Viaskin milk group (50.1+/–43.97) compared to placebo (48.20+/–56.98) in the intent-to-treat (ITT) population (Viaskin milk n=15, placebo n=5). The failed response in the ITT population may have been confounded by protocol violations in both arms regarding diet and PPI use. In the per protocol analysis (Viaskin milk n=7, placebo n=2), subjects in the Viaskin milk group demonstrated significantly lower mean eos/hpf (25.57+/– 31.19) compared to placebo (95.00+/–63.64) (p=0.038). After continuation with open label Viaskin milk for 11 months in 19 patients, 6 had <6 eos/hpf (32%) and 3 had 7–14 eos/hpf (16%), for a total response rate of 47%.³⁷

Follow up from this pilot study suggests that the treatment effect of EPIT on EoE can persist for 2 years after stopping therapy. Six of 7 patients in the complete responder (<6 eos/hpf) and partial responder (6–14 eos/hpf) groups successfully continued with 2 servings of milk a day without exacerbation of symptoms. Of these patients, one complete responder had 0 eosinophils and one partial responder had 6–14 eos/hpf on follow up esophageal biopsies.³⁸ Further investigation on efficacy and long-term data is needed to enhance our understanding of the potential role of EPIT in the treatment of EoE.

Summary

In summary, immunotherapy may have mixed effects on EoE development and disease depending on the route, dose, nature of the antigen(s), and status of the epithelium. Data is uncertain regarding improvement in EoE disease activity with aeroallergen SCIT, and SLIT may induce EoE. OIT for IgE-mediated food allergy may unmask and/or induce EoE; whereas, EPIT for non-IgE-mediated, milk-induced EoE has been investigated as a therapeutic. Different EoE outcomes are likely related to mode of administration and immune mechanism. For example, in OIT, a potentially compromised esophageal epithelial barrier is directly exposed to allergens known to cause IgE-mediated effects; whereas, in EPIT for EoE, a presumably intact skin barrier is exposed to an EoE food trigger. The presentation of EoE during immunotherapy presents several ethical dilemmas; therefore, the risks and benefits of continuing immunotherapy must be weighed for each patient. Additional studies are needed to identify the cellular and molecular mechanisms underlying the relationship between EoE and immunotherapy in order to identify those patients who are at risk for this complication.

Key Points:

• Oral immunotherapy (OIT) for food allergy and sublingual immunotherapy (SLIT) for aeroallergen sensitivity may induce EoE.

• Most patients experience disease remission upon discontinuation of OIT or SLIT.

• Aeroallergens may also trigger EoE.

• Subcutaneous immunotherapy (SCIT) and epicutaneous immunotherapy (EPIT) require further study as potential treatments for EoE.

Synopsis:

Immunotherapy is a treatment approach based on the principle of incremental allergen exposure to achieve desensitization. Recently, oral immunotherapy has been introduced as a treatment for IgE-mediated food allergy. Some patients receiving oral immunotherapy for food allergy may develop eosinophilic esophagitis. Here, we summarize the literature examining this association, its treatment and outcomes and discuss possible explanations for this clinical phenomenon. We further identify potential associations with aeroallergen sensitivity and other forms of immunotherapy including subcutaneous immunotherapy and sublingual immunotherapy. Finally, we discuss management of immunotherapy-induced EoE, and epicutaneous immunotherapy is highlighted as an area of therapeutic investigation.

Clinics Care Points.

• OIT induces esophageal eosinophilia, which may be transient and asymptomatic.

• OIT and SLIT may induce EoE.

• Most patients experience disease remission upon discontinuation of OIT or SLIT.

• Aeroallergens may trigger EoE.

• Additional studies are needed to determine whether SCIT decreases EoE disease activity in patients with aeroallergen sensitivity.

• EPIT is being investigated as a treatment for EoE.

Abbreviations used:

AEC

absolute eosinophil count

EGD

esophagogastroduodenoscopy

EGID

eosinophilic gastrointestinal disease

EoE

eosinophilic esophagitis

EPIT

epicutaneous immunotherapy

ITT

intent-to-treat

OIT

oral immunotherapy

OITIGER

oral immunotherapy-induced gastrointestinal symptoms and eosinophilic responses

PPI

proton pump inhibitor

SCIT

subcutaneous immunotherapy

SLIT

sublingual immunotherapy