Screening children for eosinophilic esophagitis: allergic and other risk factors

Eosinophilic esophagitis (EoE) is a chronic allergic inflammatory disease of the esophagus that is triggered by specific foods. EoE presents across the lifespan, with symptoms of failure to thrive, poor weight gain, feeding intolerance, food refusal, vomiting, dysphagia and food impaction [1,2]. Any patient may present with symptoms of EoE, regardless of preexisting risk factors. All patients with symptoms of esophageal dysfunction suggestive of EoE should receive screening via esophagastroduedenoscopy (EGD) with esophageal biopsies [3,4]. However, many patients with EoE develop symptoms slowly over time, and the frequency of compensatory eating behaviors is high in this population [5–7]. Therefore, it is important to consider which patient populations may be at increased risk for the development of EoE and ask targeted screening questions in order to pinpoint esophageal symptoms. Here we review epidemiologic and clinical features that may increase suspicion for EoE.

A family or personal history of preexisting allergic disease should increase the suspicion for EoE, as it has been reproducibly demonstrated in multiple cohorts that individuals with EoE have a high prevalence of allergic comorbidity [8,9]. As atopic disorders are prevalent, screening for EoE is an important consideration for general pediatricians in addition to allergists. A recent meta-analysis by González-Cervera et al examined 21 observational case-control studies of 53,542 individual EoE patients, demonstrating that EoE patients have higher rates of allergic comorbidity than the background population [10]. The odds of a personal history of atopy in EoE patients was increased, including reported rates of eczema (odds ratio [OR] 2.85, 95% confidence interval [CI] 1.87-4.34), allergic rhinitis (AR; OR 5.09, 95% CI 2.91-8.90), and asthma (OR 3.01, 95% CI 1.96-4.62). Using a large pediatric virtual birth cohort, we have measured these risk relationships in individuals revealed that history of AD (hazard ratio [HR] 3.2, 95% confidence interval [CI] 2.2-4.6), IgE-mediated food allergy (IgE-FA; HR 9.1, 95% CI 6.5-12.6), and asthma (HR 1.9, 95% CI 1.3-2.7) are independently and cumulatively associated with subsequent EoE diagnosis (Table 1) [11].

Table 1.

Effect of primary atopic diagnosis (1° Diagnosis) on likelihood of developing EoE as single diagnosis or in combination. Case-control analysis was performed on a primary care birth cohort of 130,435 children in a United States health system to determine the extent that existing allergic conditions influence the rate of subsequent EoE diagnosis. Data shown as Cox hazard ratios (HR) with 95% confidence interval (95% CI) indicating likelihood of developing EoE. From Hill DA, et al., J Allergy Clin Immunol Pract 2018, 6:1528–1533.

1° Diagnosis	EoE HR [95% CI]
AD	3.2 [2.2-4.6]
IgE-FA	9.1 [6.5-12.6]
Asthma	1.9 [1.3-2.7]
AR	2.8 [2.0-3.9]
Any 1	2.5 [1.6-3.9]
Any 2	5.6 [3.5-8.9]
Any 3	9.1 [5.2-16.0]

The relationship between IgE-FA and EoE is particularly strong—children with IgE-FA develop EoE at 9 times the rate of children without IgE-FA [12]. IgE-FA and EoE has been reported to co-occur at rates between 9.8–44% in EoE cohorts, up to as high as 68% [10,13]. Specifically, IgE-FA to milk, egg or shellfish has been associated with development of EoE [13]. Barbosa et al. studied esophageal eosinophilia in a pediatric cohort with persistent milk IgE-FA, and found that 38% had esophageal eosinophilia on biopsy with 23.5% of patients exhibiting symptoms [14]. In a cohort of adult patients with IgE-FA undergoing screening prior to oral immunotherapy (OIT), Wright et al. found 14% of patients had >15 eosinophils per high-power field (eos/hpf) [15]. These studies illustrate a growing recognition of the unique risk association between IgE-FA and EoE.

An additional concern is the risk for EoE development in IgE-FA patients on oral immunotherapy for food allergy. The risk of developing EoE has been quantitated in recent reviews as between 2.7%-5.3%, however this is likely an underestimate. Gastrointestinal side effects like abdominal pain and vomiting are common in OIT, with approximately 30% of patients reporting these symptoms which can be presenting symptoms of EoE. However, biopsies to assess for EoE were performed in a minority of studies [16,17]. This suggests that the rate of EoE is increased in OIT patients and if symptoms are persistent they should be screened with biopsy to assess for EoE.

Multiple reports have shown EoE to be associated with several non-allergic disorders including gastrointestinal, autoimmune disorders, and connective tissue diseases (Table 2) [18–20]. The association of EoE with celiac disease has been reported but is somewhat uncertain. We have recently seen an association between EoE and celiac disease in our pediatric referral population, with 3–5.6% of celiac disease patients with concurrent esophageal eosinophilia on biopsy [9,26,27]. While some investigators have seen similar associations, some studies have not observed this association [28,29]. Esophageal eosinophilia has been also noted to precede or coexist with development of inflammatory bowel disease, and the incidence of IBD in patients diagnosed with EoE in our cohort of pediatric patients is increased [9,30,31]. However, the precise clinical or mechanistic relationship of tissue eosinophilia to IBD has not been established. It is debated whether the increased association of some GI disorders with EoE results from potential bias associated with increased endoscopic screening or points to commonly dysregulated immune mechanisms.

Table 2:

Additional diagnostic considerations during the evaluation of esophageal eosinophilia. Top Pane: a number of diseases can affect the esophagus resulting in a secondary eosinophilic infiltrate and should be ruled out prior to diagnosing EoE. Bottom Pane: disorders reported to have association with a diagnosis of EoE.

Causes of non-EoE esophageal eosinophilia:
• Gastroesophageal reflux disease	[21]
• Infection(s): fungal, viral
• Drug Hypersensitivity
• Vasculitis
• Graft versus host disease
• Achalasia
• Pemphigus vegetans
Conditions associated with increased risk of eosinophilic esophagitis:
• Atopy: eczema, asthma, food allergy	[10,12]
• Inflammatory bowel disease	[16,17]
• Monogenic conditions:
○ Connective tissue disease Marfans II, Loeys-Dietz, Ehlers-Dhanlos)	[20]
○ Netherton Syndrome (SPINK5 deficiency)	[22]
○ PTEN hamartoma syndrome: 200x risk of EoE/EGID	[23]
○ Severe dermatitis, multiple allergies, and metabolic wasting (SAM), DSG1 deficiency	[24]
○ Hyper-IgE syndrome	[25]

Within some pediatric cohorts, patients with connective tissue disorders have been identified to be at significantly increased risk for EoE. The risk of EoE was found to be increased 8-fold in patients with connective tissue disorders including Marfan’s, Ehlers-Danlos, and Loeys-Dietz syndromes within one US pediatric cohort [20]. Patients within this population had typical syndromic features including characteristic facies, hypermobility, and lower BMI compared to EoE-only controls. Extra-esophageal eosinophilic gastrointestinal disease was found in 24% (10/24) in this population, making this complication unusually common in the cohort with connective tissue disorders [20]. Heart defects characteristic of connective tissue disorders were also common. Within our distinct pediatric cohort, we have seen an increase in the risk 11.0 (OR, 95% CI 4.9-24.8) of connective tissue disease in our EoE cohort once data was adjusted for age and gender. However, when some population of connective tissue disorder disease patients are retrospectively reviewed, EoE does not emerge as a major comorbidity in these patients [32]. It can be hypothesized that this is due to the studies’ focus on children and adults and that more evidence will emerge about this link will emerge as familiarity with these diagnoses increases among providers.

Of particular importance for the practicing immunologist is the association of EoE with hyper-IgE syndrome caused by STAT3 mutations (AD-HIES) [25]. 60% of 70 patients in a large American cohort of AD-HIES syndrome had chronic GI complaints. Of these, 23 patients underwent esophagogastroduodenoscopy, demonstrating eosinophilic inflammation in 65% of these patients. This suggests that secondary eosinophilic esophagitis is a significant consideration in these patients. EoE has additionally been described in case reports of Common Variable Immunodeficiency [33,34].

In addition, we observed a higher prevalence of Autism Spectrum Disorder (ASD) within our cohort of EoE patients; the rate of ASD in children with EoE is 7.5%, compared to 1.9% in those without EoE (OR 4.2, 95% CI 2.9-6.0, P<0.0001) [9]. While the etiology of this relationship remains unknown, the often-severe adverse feeding behaviors that can be characteristic of ASD may in part be due to underlying and potentially undiagnosed esophageal disease. These findings support a recommendation screen for EoE in patients with ASD and unexplained feeding dysfunction and highlights a potential role for nutritionists and occupational therapists in screening for EoE. These specialists can play a pivotal role in recognizing when abnormal feeding behaviors may be indicative of esophageal or other GI dysfunction. Future research efforts are needed to verify the extent of this association on a population scale. Delineating specific symptoms indicative of EoE as opposed to a behavioral feeding disorder in ASD will be paramount to providing more specific evidence-based clinical recommendations.

Regardless of preexisting risk factors, EoE should be suspected in any patient with a consistent clinical presentation. Characteristic symptoms that suggest a diagnosis of EoE are those of chronic esophageal dysfunction Symptoms vary with patient age. Failure to thrive and feeding problems are seen more commonly in younger children whereas older adolescents present more frequently with dysphagia, odynophagia and food impaction. Patient-reported outcome tools have been developed for EoE, but their utility for diagnostic purposes is limited. This is because it has been observed in randomized control trials that esophageal eosinophilia may be present without esophageal symptoms and vice versa [35]. As a consequence, diagnostic application of patient-reported outcome tools has had limited sensitivity and specificity, but have had some success in following patients longitudinally for clinical trials. The PEESS v2.0 is a pediatric-specific, validated questionnaire to assess patient symptoms [36]. The survey has been validated for use in children (ages 8-18) and for completion by a parent-proxy (ages 2–18). Scoring for dysphagia in PEESS is highly correlated to increased EoE disease activity on biopsy. In adult populations, the Straumann dysphagia instrument [5], the Dysphagia Symptom Questionnaire [37], and the EEsAI PRO instrument [38] have been validated and incorporate slightly different aspects of EoE presentation into their symptom questionnaires. There is interest in refining these scales to improve their ability to be used to follow EoE patients longitudinally over time.

Finally, it is important for all physicians to be aware of recent updates in the diagnostic criteria for eosinophilic esophagitis [3,4]. Previously, response of esophageal eosinophilia to proton pump inhibitor (PPI) therapy was considered diagnostic for gastroesophageal reflux disease (GERD), a distinct disease entity. As such, the 2007 diagnostic guidelines required a diagnostic trial of high-dose proton pump inhibitor (PPI) to rule out GERD [39]. There is a growing recognition that approximately 50% of patients with EoE have decreased mucosal inflammation with PPI therapy, indicating that PPI is more correctly considered a therapy for EoE. Updated diagnostic criteria from 2017 indicate that PPI trial prior to endoscopy is not required to make a diagnosis of EoE [3]. All patients with greater than 15 eos/hpf on esophageal biopsy meet criteria for EoE diagnosis, provided that there are symptoms of esophageal dysfunction and non-EoE etiologies of esophageal eosinophilia have been excluded (Table 3). The decision to initiate or continue PPI therapy, therefore, should be guided by patient symptoms.

Table 3.

Summary of updated EoE diagnostic criteria [3,4]

1. Symptoms of esophageal dysfunction

○ Concomitant atopic conditions should increase suspicion for EoE

○ Endoscopic findings characteristic of EoE should increase suspicion

2. Biopsy findings of 15 eosinophils per high-power field (approximately 60 eosinophils/mm2), with tissue eosinophilia isolated to the esophagus

3. Other causes of esophageal eosinophilia have been ruled out, for example: eosinophilic gastrointestinal disease with esophageal involvement; achalasia and other disorders of esophageal dysmotility; Hypereosinophilic syndrome; Crohn disease with esophageal involvement; infections (fungal, viral); connective tissue disorders; dermatologic conditions with esophageal involvement (ie: pemphigus); drug hypersensitivity reactions; pill esophagitis; graft vs host disease; Mendelian disorders (Marfan syndrome type II, hyper-IgE syndrome, PTEN hamartoma tumor syndrome, Netherton syndrome, severe atopy metabolic wasting syndrome)

In our experience, in pediatric patients under the age of 12, GERD symptoms can overlap with symptoms of EoE. Therefore, many younger pediatric patients benefit from a trial of PPI both to determine the extent to which GERD may affect their EoE and also to determine if this improves symptoms. Some children may experience quick improvement without symptom recurrence, wean off of PPI successfully and require no further intervention. However, if patients cannot stop taking PPI without recurrent symptoms, then esophagastroduedenoscopy can still be performed when the patient is on high-dose PPI. However, if the eosinophil count is under 15 eosinophils/hpf while on high-dose PPI in a symptomatic patient, it does not rule out a diagnosis of EoE which has been treated by PPI.

In summary, EoE should be suspected in any patient with chronic symptoms of esophageal dysfunction. In addition, screening for a personal history of atopy, with particular emphasis on current or prior IgE-mediated food allergy, can aid in identification of at-risk individuals. Esophageal eosinophilia has been associated with a number of syndromes (Table 2). In these patients, evidence suggests that the pretest probability for EoE may be higher and when esophageal symptoms are present, and referral for EoE screening by EGD with biopsy may help to curtail diagnostic odyssey and promote use of correct therapy for patients.

Abbreviations:

EoE

eosinophilic esophagitis

AR

allergic rhinitis

AD

atopic dermatitis

IgE

immunoglobulin E

HR

hazard ratio

ASD

Autism Spectrum Disorder

PPI

proton pump inhibitor

GERD

gastroesophageal reflux disease

PPI-REE

proton pump inhibitor-responsive esophageal eosinophilia