Eosinophilic esophagitis: update on management and controversiesBMJ 2017; 359 doi: https://doi.org/10.1136/bmj.j4482 (Published 13 November 2017) Cite this as: BMJ 2017;359:j4482
- Joan W Chen, clinical assistant professor,
- John Y Kao, associate professor
- 1Division of Gastroenterology and Hepatology, Department of Internal Medicine, University of Michigan, Ann Arbor, MI 48109-5362, USA
- Correspondence to: J W Chen, 3912 Taubman Center, 1500 E. Medical Center Drive, SPC 5362, Ann Arbor, MI 48109-5362, USA
Eosinophilic esophagitis is a chronic allergen driven immune mediated disease that is increasingly recognized as a leading cause of dysphagia and foregut symptoms in children and adults. Much knowledge has been gained in recent years on the genetic and environmental risk factors for this disease, the associated inflammatory milieu, and the long term complications from esophageal remodeling. In this review we will highlight recent progress made in research into this disease, focusing on adults. We will discuss ongoing efforts to develop a minimally invasive technique that may obviate the need for repeated endoscopic assessment of disease activity. Moreover, we will review studies using novel tools such as mucosal impedance and functional lumen imaging as potential surrogate markers for mucosal integrity and esophageal remodeling. With regard to the treatment of eosinophilic inflammation, we will discuss the controversies surrounding responsiveness to proton pump inhibitors in some patients. Therapeutic trials continue to support the use of topical glucocorticoids and empiric food elimination diets as first line treatments. We will discuss ongoing efforts to optimize the elimination diet protocol to decrease the level and duration of food restrictions. Looking ahead, our growing knowledge on the pathogenesis of eosinophilic esophagitis has enabled further advancement of promising targeted biologic therapies.
Eosinophilic esophagitis is a chronic immune mediated condition characterized clinically by symptoms related to esophageal dysfunction and histologically by eosinophilic predominant inflammation.1 Although it has not been associated with an increased risk of mortality or cancer, the chronic progressive nature of the disease has a negative impact on patients’ quality of life.2 Since its recognition as a clinical entity in the 1990s,34 the disease has evolved from sporadic case reports to become a widely recognized cause of esophageal morbidity.56 However, despite growing efforts to elucidate its pathogenesis and the increasing number of well designed clinical trials, the practical management of eosinophilic esophagitis is often challenging because of controversies related to therapeutic endpoints and the need for long term treatments. To date, three main consensus recommendations for management have been published.178 The most recent of these guidelines was published in 2013, and most of the recommendations were “conditional” rather than “strong,” because of the relatively low quality of evidence.
We conducted a thorough review of the literature, with an emphasis on studies published since the most recent consensus guideline and focusing on adults with eosinophilic esophagitis. Our review includes recent discoveries on pathogenic mechanisms, genetic and environmental risk factors, advances and controversies in the development of a minimally invasive diagnostic biomarker, techniques to quantify fibrostenosis and mucosal integrity in eosinophilic esophagitis, novel treatment options including biologic agents, and optimization of the dietary elimination therapy. Finally, we propose key research questions to be addressed in future translational research and therapeutic trials.
Sources and selection criteria
We searched PubMed up to April 2017 to identify English language articles published in peer reviewed journals using a list of search terms that can be found in the supplemental material (see appendix at end of document for details). Reference lists from selected manuscripts were also used to identify additional relevant articles. We prioritized recent prospective and randomized controlled trials (RCTs), although laboratory studies, population based and observational studies, meta-analyses, and systematic reviews are included whenever prospective controlled trials were unavailable. The search terms used did not limit the results to adults or children because many studies included both age groups. However, the review focuses on adults because pediatric specific data are limited. Data on pediatric specific eosinophilic esophagitis, when included, will be specified.
Incidence and prevalence
The incidence and prevalence of eosinophilic esophagitis have risen over the past two decades.168910 In a report from Olmstead County, Minnesota, USA, no cases were documented before 1990, and the incidence rose from 0.35 cases/100 000 people to 9.5 cases/100 000 people over a 15 year period.6 Increased incidence has also been reported in Switzerland,5 the Netherlands,11 and Denmark.12 The reasons for this increase are poorly understood, and increased recognition is probably not the only cause.56
Using a large scale administrative claims database from the United States, a recent study reported about 152 000 cases of eosinophilic esophagitis in the US, with a prevalence of 39/100 000 to 153/100 000, depending on the case definition used.13 Pooled data from a systematic review showed that the sample size weighted average was 0.03% in two population based studies and 2.8% in seven studies evaluating symptomatic patients.14 There are also regional differences in prevalence; for example, prevalence in Western countries, including the US and Europe, is estimated at 0.4-1%,151617 whereas in Japan it is estimated to be much lower (0.01%).1819 Three to four times more males are affected than females, and the disease is more common in white people than in other ethnic groups.202122 A recent meta-analysis that extracted adult only or pediatric only prevalence data reported a higher prevalence in adults than in children (43.4/100 000 (95% confidence interval 22.5 to 71.2) v 29.5/100 000 (17.5 to 44.7).15
Animal and human studies have shown that an allergen mediated inflammatory process is the key mechanism of pathogenesis. Between 50% and 80% of patients have concurrent allergic diseases including food allergies, allergic rhinitis or sinusitis, atopic dermatitis, and asthma.10212324252627 However, unlike asthma or other common immune mediated diseases, eosinophilic esophagitis does not seem to be a classic IgE mediated immune response. Rather, it is a T helper type 2 cell (Th2) predominant inflammatory process triggered by allergens.282930 In an experimental animal model, eosinophilic esophagitis can be induced directly by delivering cytokines produced by Th2 cells.31 In mice, deletion of central downstream mediators in the Th2 allergic cascade, such as interleukin 4 (IL-4), IL-5, and IL-13, protects against the development of eosinophilic esophagitis,3132 and overexpression of IL-5 leads to an eosinophilic esophagitis phenotype.33
The Th2 inflammatory cascade
Figure 1⇓ illustrates a proposed molecular pathogenic pathway of eosinophilic esophagitis. An increase in thymic stromal lymphopoietin (TSLP) is thought to induce Th2-type immune mediated inflammation. Genetic gain-of-function mutations of genes encoding TSLP and the TSLP receptor have been shown to favor the development of eosinophilic esophagitis, and TSLP expression is significantly higher in patients with active disease.343536 Th2 lymphocytes subsequently produce a combination of type 2 cytokines, including IL-4, IL-5, and IL-13.37 IL-4 promotes the differentiation of T cells into Th2 cells and has been linked to other atopic diseases.38 IL-5 is essential for the proliferation and maturation of eosinophils,39 and IL-13 promotes eosinophilic infiltration by stimulating esophageal epithelial cells to produce eotaxin-3, a potent chemokine that recruits and activates eosinophils.40 In addition, IL-13 has been associated with dysfunction of the epithelial barrier.29
Eosinophils, a hallmark cell type found in eosinophilic esophagitis, are recruited to the esophagus by eotaxin-3 and chemoattractant proteins such as CRTH2 (prostaglandin D2 receptor 2).414243 Once eosinophils are recruited to the esophagus, they are activated to release granules containing major basic protein, eosinophil derived neurotoxin, eosinophil cationic protein, and eosinophil peroxidase,4445 in addition to several cytokines. Together, they facilitate inflammation,44464748 increase smooth muscle activation,4950 and trigger degranulation of mast cells and basophils.51 Mast cells have also been implicated; their numbers are increased in patients with eosinophilic esophagitis compared with controls and those with gastroesophageal reflux disease (GERD).5253545556575859 Mast cells may promote inflammation and fibrosis in eosinophilic esophagitis by secreting transforming growth factor β (TGF-β),50 which results in epithelial proliferation, epithelial-mesenchymal transition, deposition of collagen, and fibrosis of the esophagus.60616263 Recently, basophils and invariant natural killer (iNK) cells have also been linked to the disease,6465 but their roles are not well understood.
Additional insights from genetic studies
Recent genetic studies have helped gain additional insights into the molecular pathogenesis of eosinophilic esophagitis. Genome-wide RNA profiling of esophageal biopsies found that CCL-26, which encodes eotaxin-3, a potent chemoattractant of eosinophils stimulated by IL-13, was overexpressed about 50-fold in patients with eosinophilic esophagitis compared with normal controls or those with GERD.66 Another gene of interest, the gene encoding filaggrin, is crucial in skin barrier function.2937 It is downregulated by IL-13, leading to the disruption of esophageal wall integrity and an increase in antigen exposure.37 Other genes that are potentially important in the pathogenesis of eosinophilic esophagitis have been identified through recent genome-wide array studies (GWAS),3467 but further studies are needed to improve our understanding of their functional importance.
Accumulative evidence now shows that the pathogenesis of eosinophilic esophagitis is multifactorial, and that it probably arises from an allergic response to environmental factors in genetically susceptible individuals.7
Genetic epidemiologic studies of families suggest that genetics plays a role in susceptibility, although the disease does not follow the classic Mendelian inheritance pattern.968 Monozygotic twin studies show a high concordance for the trait, estimated at 41%69; 21% of people with an affected dizygotic twin developed the disease, whereas this figure was only 2.4% in the nuclear family cohort. This supports the additional contribution of environmental factors to phenotypic variance. In another study, the sibling risk ratio—the ratio of risk in people with an affected sibling compared with the disease prevalence in the general population—was 60.69 Recurrence risk ratios ranged from 10 to 64, depending on the family relationship—it was highest in brothers, followed by fathers, and lower in sisters and mothers.69
Recent GWAS have suggested that eosinophilic esophagitis, atopic diseases, and autoimmune conditions have a shared genetic etiology.3467 The Utah Population Database (covering >85% of Utah’s population) was used to determine associations between specific autoimmune diseases in probands diagnosed with eosinophilic esophagitis and their extended family members.70 First degree relatives of people with eosinophilic esophagitis had an increased risk of several autoimmune conditions including celiac disease, Crohn’s disease, ulcerative colitis, rheumatoid arthritis, IgA deficiency, common variable immune deficiency, multiple sclerosis, and Hashimoto’s thyroiditis. In addition, probands and their families with ulcerative colitis, systemic sclerosis, and multiple sclerosis had an increased risk of having eosinophilic esophagitis.
Early life risks factors
Several early life perturbations increase the risk of developing eosinophilic esophagitis. One study found an association between early use of antibiotics and the development of esophageal eosinophilia (odds ratio 6.1, 1.7 to 20.8).71 Another study found a link between the development of esophageal eosinophilia and cesarean delivery and antibiotic use (3.2, 1.2 to 8.6 and 3.6, 1.3 to 10.1 respectively).72 Finally, an inverse association has been reported between exposure to smoking in the first year of life and esophageal eosinophilia (0.48, 0.23 to 1.02).73 A causal association between these early life risk factors and esophageal eosinophilia has not been proved, but one theory is that early exposure can lead to altered gut microbiota (dysbiosis), which may have an effect on normal immune development.74
The human microbiome has been the subject of intense research interest in recent years; however, there is still a paucity of data on the microbiome of the esophagus. The composition of the intestinal microflora and the resulting immunologic response are important for the induction of tolerance. In dysbiosis, the metabolic profile shifts towards a more favorable environment for pathogenic constituents, disrupting immune regulatory systems and leading to a state of immune intolerance.75 In esophageal eosinophilia, the absence of beneficial bacteria may lead to a decrease in the induction of regulatory T cells that suppress proinflammatory effector cells (such as eosinophils), as well as disruption of the gastrointestinal barrier function. There is some evidence to suggest that the alteration in gut microbiota in early life (for example, from the use of antibiotics) might trigger an abnormal immune response that persists throughout life. Two recent studies have shown a difference in the load and composition of esophageal microbiota in patients with esophageal eosinophilia compared with controls.7677 With advances in high throughput DNA sequencing technologies and bioinformatics tools, disease associated esophageal microbiota signatures may emerge that will enable the development of microbiota biomarkers that can help define clinical phenotypes.
Helicobacter pylori has been inversely associated with allergic diseases such as asthma, allergic rhinitis, atopic dermatitis, and inflammatory bowel disease.787980 The specific mechanism leading to a protective role for H pylori in allergic diseases is unknown but is probably related to systemic immunomodulation induced by the infection.
The inverse association between H pylori and esophageal eosinophilia was first observed in a population based study from Sweden,17 in which the infection was found to be an independent negative predictor for esophageal eosinophilia (odds ratio 0.41, 0.19 to 0.92). In a cross sectional study using a pathology database of 165 000 patients with both esophageal and gastric biopsies, the presence of esophageal eosinophilia was inversely associated with H pylori infection (0.77, 95% CI 0.69 to 0.87).81 The odds appear to reduce further as levels of esophageal eosinophilia increase. Two case-control studies have also supported the inverse association between H pylori and eosinophilic esophagitis.8283 Because the prevalence of esophageal eosinophilia and H pylori infection both vary widely by patient ethnicity,83H pylori may be partly responsible for the observed ethnic distribution of the disease.
Currently, there are two hypotheses for the potential protective effect of H pylori on esophageal eosinophilia. The first is the hygiene hypothesis, which attributes the rise in allergic diseases to improved sanitation and reductions in childhood infections, including H pylori infection. The other involves the observation that H pylori may drive a Th1 mediated allergic response, shifting the Th1–Th2 balance, and thus “protecting” from an allergic Th2 response in patients with eosinophilic esophagitis.
Presentation and diagnosis
Adults and children with eosinophilic esophagitis have different clinical presentations.84858687 Children commonly present with difficulty feeding, vomiting, abdominal pain, or failure to thrive, whereas adults typically have symptoms of dysphagia, heartburn, chest pain, and food impactions.208488 This difference is thought to be, at least in part, related to gradual esophageal tissue remodeling arising from unabated inflammation.89909192 An inflammatory phenotype—endoscopic features including white exudates, linear furrows, and edema—is thought to eventually evolve as a result of longstanding inflammation into a fibrostenotic phenotype, with fibrotic features on endoscopy such as rings, strictures, and a narrow caliber esophagus.899394
Monitoring of disease activity and treatment endpoint
Potential treatment endpoints in eosinophilic esophagitis include improvements in clinical symptoms, esophageal eosinophilic inflammation, and severity of endoscopic features. The use of treatment endpoints varies considerably in the literature, and this variability has hindered the interpretability and comparability of clinical trials.95 Owing to the relative ease in identifying and interpreting esophageal eosinophilia, most studies have relied on histology to assess therapeutic response. However, the specific threshold for a reduction in eosinophilia to define treatment efficacy remains uncertain, and <20, <15, <10, <5 eosinophils/high power field (HPF) have all been used in the literature. Both mean eosinophil and peak eosinophil densities have also been used. The absence of subepithelial tissue in routine biopsies for assessment of esophageal remodeling further hampers the utility of histologic examination.96 Recent validation of patient reported outcome (PRO) instruments—which combine an endoscopic grading system, histologic scoring system, and surrogate measurements of esophageal remodeling and tissue integrity—have led the way in establishing more consistent and meaningful treatment endpoints.
Clinical assessment: PRO instruments
Symptom assessment can be challenging because most patients have chronic progressive symptoms and may have developed adaptive behaviors such as avoidance of specific food textures, excessive mastication, or increased meal times to avoid dysphagia and food impaction. In addition, dysphagia and food impaction may occur as sporadic events in many patients and may not be captured in therapeutic studies of short duration. PRO instruments have recently been developed and validated in adults, including the eosinophilic esophagitis activity index (EEsAI) and dysphagia symptom questionnaire (DSQ).97 The EEsAI is composed of seven PRO items that assess esophageal symptoms over a seven day recall period. The instrument accounts for behavioral adaptations such as avoidance of specific food textures and meal time length in adults with eosinophilic esophagitis.98 However, in a prospective, observational study of 269 adults with eosinophilic esophagitis, the EEsAI score alone was inadequate in the detection of patients with endoscopic, histologic, and combined remission (area under the curve values of 0.67, 0.60, and 0.67, respectively).99 The authors cautioned that assessment of disease activity on the basis of symptoms alone is probably inadequate.
The presence of endoscopic features is not required for the diagnosis of eosinophilic esophagitis, but the endoscopic presence of edema, rings, exudates, or strictures increases the clinical likelihood of histologic esophageal eosinophilia. Prospective studies have identified endoscopic features in 93% of patients with eosinophilic esophagitis according to a meta-analysis that pooled data from 15 prospective studies (N=378 patients).100 A classification and grading system to assess the endoscopic findings in the disease—the endoscopic reference score (EREFS)—was proposed in 2013 in an effort to standardize endoscopic assessment (figs 2 and 3⇓).101 Interobserver agreement was moderate for rings, white exudates, furrows, and strictures, and intraobserver agreement was moderate to substantial for rings, furrows, white exudates, and strictures in a study that enrolled four experts and four trainee endoscopists to assess the endoscopic images of 30 eosinophilic esophagitis patients.102
The diagnostic utility of the EREFS system was investigated in a recent prospective study that compared EREF scores between 67 patients with eosinophilic esophagitis and 144 controls103; the mean total EREF score was 3.88 in patients and 0.42 in controls (P<0.001, area under the curve (AUC) 0.934). After treatment, the mean EREF score in patients decreased from 3.88 to 2.01 (P<0.001). This change was more prominent in patients with a histologic response than in non-responders. The study suggests that endoscopic EREFS evaluation can be used to identify those with eosinophilic esophagitis and to monitor treatment response. However, another recent retrospective study of 69 patients with eosinophilic esophagitis found that EREF scores correlated only weakly to moderately with peak eosinophil count.104 Therefore the system’s predictive value of disease activity may be insufficient for clinical use, and biopsies remain indispensable for the assessment of disease activity.
Eosinophilic esophagitis is associated with characteristic esophageal histopathology, including increased intraepithelial eosinophils, eosinophil abscesses, basal cell hyperplasia, and lamina propria fibrosis.105106 Besides increased eosinophil density (≥15 eosinophils/HPF), the pathologic changes are characteristic but not pathognomonic. A recent study found that even pathology trainees can accurately diagnose eosinophilic esophagitis by histologic examination.107 A histologic scoring system for esophageal biopsies that incorporates evaluation and scoring based on eight features: eosinophil density, basal zone hyperplasia, eosinophil abscesses, eosinophil surface layering, dilated intracellular spaces, surface epithelial alteration, dyskeratotic epithelial cells, and lamina propria fibrosis has recently been proposed.108 This method objectively assesses histologic changes in the esophagus beyond numbers of eosinophils, but the clinical utility of this scoring system has not been validated.
The quest for a non-invasive biomarker for diagnosis and monitoring of disease activity is an ongoing effort of high priority. Biomarkers investigated either involved immunohistochemistry on esophageal biopsy samples58109110111 or clinical immunology techniques on serum samples.112113114 To date, no single marker in blood or esophageal tissue has proved reliable for diagnosis.115116 Recently, the focus has shifted to the development of a panel of biomarkers. A molecular diagnostic test that used a 94 gene quantitative polymerase chain reaction (PCR) array and a scoring system was highly sensitive and specific in identifying active eosinophilic esophagitis. It also distinguished between patients with eosinophilic esophagitis who were in remission and controls, as well as those exposed to swallowed glucocorticoids.117 Another study that investigated the diagnostic value of a 94 gene expression score using PCR of prospectively collected esophageal tissue in adults found that the score had high diagnostic utility for distinguishing between patients with and without eosinophilic esophagitis, with an AUC of 0.927.118 These techniques would not eliminate the need for endoscopy with esophageal tissue collection but could be useful for patients with diagnostic uncertainty on the basis of conventional testing.
Office based testing
Office based testing of disease activity for diagnosis or assessment of treatment response has been of great interest. The esophageal string test44 and cytosponge119 are examples of minimally invasive clinical devices that have been tested in small prospective trials to collect esophageal samples. Multicenter studies involving these two tests are currently under way.
Measurement of luminal distensibility
The dominant symptoms in adults—dysphagia and food impaction—are often more relevant to tissue remodeling and fibrosis rather than active eosinophilic inflammation. Fibrosis of the subepithelial space has been demonstrated in both pediatric and adult cohorts90120; however, in standard esophageal biopsies, only the superficial epithelial layer of the esophagus is examined. The introduction of high resolution impedance planimetry has enabled the assessment of tissue remodeling and its effect on mechanical properties of the esophagus. The functional luminal imaging probe (FLIP) uses high resolution impedance planimetry to render a three dimensional approximation of intraluminal esophageal anatomy during volumetric distention. One study prospectively analyzed the pressure geometric properties of the esophagus in 33 patients with eosinophilic esophagitis and 15 controls using FLIP. It showed that distensibility, defined by the change in the narrowest measurable cross sectional area within the distal esophagus versus intraluminal pressure, was significantly reduced in patients compared with controls independent of mucosal eosinophil count.121 In a follow-up study, 70 patients with eosinophilic esophagitis and previous food impactions showed significantly reduced esophageal distensibility as measured by FLIP. In addition, reduced esophageal distensibility predicted risk of food impaction and the need for esophageal dilatation during the prospective follow-up period.122 Another cross sectional study of 72 adults with eosinophilic esophagitis showed an inverse association between endoscopic ring severity and esophageal distensibility as measured by FLIP.123 No association between distensibility and inflammation was seen, highlighting the dissociation between the fibrostenotic and inflammatory features of eosinophilic esophagitis. Figure 4⇓ shows sample FLIP images and distensibility according to ring severity (grades 0-3). As such, FLIP may be a useful tool for measuring esophageal remodeling and stratifying the risk of food impaction in patients with eosinophilic esophagitis.
Measurement of mucosal integrity
Mucosal integrity in eosinophilic esophagitis has been measured using mucosal impedance in 11 patients with eosinophilic esophagitis and 11 healthy controls using 24 hour pH impedance monitoring in a cross sectional study.124 A decreased baseline impedance value was found throughout the esophagus in patients but not in the controls. A follow-up prospective study by the same group that used electrical tissue impedance spectroscopy to compare esophageal mucosa integrity between 16 patients with eosinophilic esophagitis and 11 controls showed that electrical tissue impedance and transepithelial electrical resistance were reduced in patients.125 This suggests that esophageal mucosal integrity is impaired in eosinophilic esophagitis, potentially leading to antigen exposure.
Mucosal impedance was measured at 2 cm, 5 cm, and 10 cm above the squamocolumnar junction in a cross sectional study that enrolled 61 patients with erosive esophagitis, 81 with non-erosive reflux disease, 18 with achalasia, 15 with eosinophilic esophagitis, and 93 controls. Mucosal impedance was significantly lower (P<0.001) in patients with GERD or eosinophilic esophagitis than in patients without GERD or patients with achalasia. Furthermore, patients with GERD had lower mucosal impedance closer to the squamocolumnar junction that increased proximally (median of 1427 Ω, 2013 Ω, 3236 Ω at 2 cm, 5 cm, and 10 cm above the squamocolumnar junction, respectively), whereas patients with eosinophilic esophagitis had low mucosal impedance along the length of the esophagus (median 123 Ω, 1259 Ω, 1365 Ω at 2 cm, 5 cm, and 10 cm, respectively).126 Finally, in a prospective study, 15 patients with eosinophilic esophagitis underwent esophagogastroduodenoscopy before and after an eight week course of twice daily swallowed fluticasone and esophageal mucosal impedance was measured by esophageal electrical tissue impedance in vivo during endoscopy and transepithelial electrical resistance and molecular flux in Ussing chambers using biopsy specimens. This study found that mucosal impedance increased substantially (P<0.01 for electrical tissue impedance and transepithelial electrical resistance, and P<0.05 for transepithelial molecular flux) after treatment, indicating a restoration of mucosal integrity with topical corticosteroids.127
Proton pump inhibitors (PPIs)
According to recent practice guidelines, a trial of acid suppression is required to fulfill the diagnostic criteria of eosinophilic esophagitis.17 Patients with significant improvement of symptoms and esophageal eosinophilia after an eight week course of at least a moderate PPI dose have either GERD or PPI responsive esophageal eosinophilia (PPI-REE).128129 PPI-REE was first described in 2006.130 Since that time, multiple prospective and retrospective adult and pediatric studies have shown that at least a third of patients with esophageal eosinophilia will respond to PPI monotherapy,7129131132133134135136 and that the response to PPIs is not reliably predicted by pH monitoring, indicating a therapeutic mechanism separate from acid suppression in PPI-REE.134
The mechanism behind PPI responsiveness in PPI-REE is not fully understood. However, it has been postulated that PPI may have anti-inflammatory properties in eosinophilic esophagitis,137 because in vitro and in vivo models have demonstrated antioxidant and anti-inflammatory effects of PPIs.138 Another proposed mechanism involves repair of a leaky mucosa. In patients with PPI-REE and not eosinophilic esophagitis, PPIs partially restored mucosal integrity, thus preventing the influx of food allergens through the mucosa.125 A third potential mechanism involves blockage of Th2 cytokine stimulated esophageal secretion of eotaxin-3 by PPIs.38139
Whether PPI-REE represents GERD, eosinophilic esophagitis, or neither remains an area of controversy, although recent studies indicate that PPI-REE and eosinophilic esophagitis have important similarities. Separate studies using immunohistochemistry to quantify esophageal tissue eosinophil associated molecular markers showed that PPI-REE and eosinophilic esophagitis have similar alterations in tissue molecular markers.57140 PPI-REE and eosinophilic esophagitis have also been shown to share similar genetic alterations based on transcriptome analysis using the 94 gene eosinophilic esophagitis diagnostic panel or a whole genome approach.137141 PPI therapy reverses both molecular and genetic alterations in PPI-REE, similar to the effect of topical steroids in eosinophilic esophagitis, and early studies have shown that patients with PPI-REE respond to diet and topical steroid therapy in a similar way to those with eosinophilic esophagitis.142143 These pieces of evidence support the notion that PPI-REE probably shares the pathogenic allergic mechanisms that underlie eosinophilic esophagitis.
Sustained efficacy of maintenance PPI therapy in PPI-REE has been shown in a prospective study that included 40 patients with PPI-REE; over 80% of patients with PPI-REE remained in histologic remission on reduced dose daily maintenance PPIs.144 Adverse events from the short term use of PPIs were mild, including hypertransaminasemia and esophageal candidiasis in two patients on high dose PPIs. Although long term data are unavailable, in light of recent concerns regarding complications from chronic use of PPIs, maintenance therapy at the lowest effective dose to maintain disease remission is currently recommended.
Diet modifications have steadily become accepted as first line treatment for patients with eosinophilic esophagitis because of their high efficacy, low cost, and safety profile. Various dietary treatment modalities have been proposed and can be separated into three main categories: the elemental diet, empiric elimination diets, and elimination diet guided by allergy testing.
The elemental formula, which consists of free amino acids devoid of proteins or potential dietary antigens, has been tested in eosinophilic esophagitis. The efficacy of this diet in achieving clinical and histologic remission of disease was first shown in 1995. In this study, normalization of esophageal histology and complete clinical remission was seen in eight of 10 children with eosinophilic esophagitis after six weeks of the diet; near normalization was seen in the remaining two children.145 The elemental diet has since been shown to induce clinical and histologic improvement in a large (N=51) prospective pediatric trial (median esophageal eosinophils per high power field decreased from 33.7 to 1.0; P<0.001) and histologic remission in an adult prospective trial (N=18, esophageal eosinophilia decreased from 54 to 10 per high power field; P=0.0006).146147 Overall, effectiveness was 90.8% (84.7% to 95.5%, I2=52.3%) in both children and adults in a recent meta-analysis that included 13 retrospective studies, prospective studies, and case reports and a total of 429 patients with eosinophilic esophagitis (411 children and 18 adults).148 However, this treatment method is generally unsustainable because of its poor palatability, the negative impact on quality of life related to the avoidance of all table foods, and the high cost of the formula.
Six food elimination diet
With the clinical unfeasibility of the elemental diet, an empiric six food elimination diet (SFED) was first tested in children with eosinophilic esophagitis in 2006.149 The SFED consists of the elimination of the top six food groups most commonly associated with food allergy (cow’s milk, wheat, egg, soy, peanut/tree nut, and fish/seafood) and subsequent sequential reintroduction with histologic examination to monitor response.150 The SFED led to clinical and histologic remission in 74% of children. Similar results were seen in all age groups.148 A recent meta-analysis showed a homogeneous (I2=0%) histologic remission rate of 72% (66% to 78%) in both children and adult patients (75 children and 122 adults).148 No adverse events were reported from the SFED; however, the high level of dietary restriction and number of endoscopies during the reintroduction phase often result in patient refusal or low compliance with this diet.
Four food elimination diet
The fact that only one or two foods have been identified as triggers for eosinophilic esophagitis in most children and adults indicates that the level of dietary restriction in SFED may be unnecessary in many cases. The top four causative foods identified are cow’s milk, wheat, egg, and soy/legumes. Consequently, an elimination diet including the top four foods—the four food elimination diet (FFED)—was proposed. A prospective multicenter trial in 52 adult patients showed a 54% remission rate with the FFED.15 A large proportion of non-responders to the FFED (31%) achieved remission after broadening to the SFED, resulting in a 72% overall efficacy, similar to that expected for empiric SFED. The benefits of FFED include a less restrictive diet, better acceptance by patients, and potentially a shorter study time with fewer endoscopies needed to identify food triggers. Table 1⇓ summarizes studies of six and four food elimination diets.
Step-up elimination diet
A step-up approach, eliminating first the top two foods (milk and wheat), then top four, then top six in non-responders has recently been proposed in an abstract.161 A two food elimination diet achieved remission in 40% of patients, and remission rates increased to previously known rates with subsequent FFED and SFED in non-responders. Compared to starting with an SFED, this step-up strategy enabled diagnostic time to be reduced by 35%, prevented unnecessary dietary restriction, and reduced endoscopic procedures by 25%.
Once a food trigger that is responsible for inducing esophageal eosinophilia in an individual patient is identified, long term avoidance is advisable to maintain drug-free disease remission. Studies evaluating the short term efficacy of a maintenance avoidance diet have shown that patients who strictly avoided food(s) known to trigger disease remained in clinical and histologic remission for three years.153154159
Allergy testing directed elimination diet
Although elimination diets are successful, the need for multiple endoscopies after each food is reintroduced makes this management option untenable to many. A non-invasive, rapid investigation that can accurately predict food triggers is needed. However, several recent studies have shown that the skin prick test, patch test, and serum food antigen-specific IgE testing are not useful in identifying food triggers. The reasons for this are not fully understood but may be partly due to the involvement of a delayed type, non-IgE-mediated hypersensitivity reaction.
Skin prick tests involve the application of a droplet of commercially prepared antigen to the patient’s skin and the use of a hand held lancet to puncture the epidermis and cause subcutaneous exposure. Patients are observed for a specified time interval and characteristic changes on the skin are assessed. A study found that only one in 20 patients with eosinophilic esophagitis had food triggers correctly identified by this test,162 and another also showed that the test performed poorly, with a positive predictive value of 13% for the identification of eosinophilic esophagitis food triggers.153 A further study found that positive skin prick test results were not concordant with confirmed food triggers and that patients who test positive or negative for a given food antigen are equally likely to relapse on food reintroduction.154 On the basis of these studies, an allergy testing directed elimination diet is not currently recommended.
Topical corticosteroids have been shown to be effective in children and adults and are considered the other first line treatment option. Table 2⇓ summarizes RCTs that have investigated the efficacy of topical corticosteroids, including topical fluticasone and budesonide, in eosinophilic esophagitis. Oral viscous budesonide has been shown to have a greater mucosal contact time than the nebulized solution.167
A recent systematic review and meta-analysis of placebo controlled RCTs determining the efficacy of topical steroids174 included five RCTs that looked at 85 adults and 89 children in total.163165166168170 Overall, topical steroids, as compared with placebo, were effective in inducing complete histologic remission (odds ratio 20.81, 7.03 to 61.63). Complete or partial histologic response with topical steroids was 57.8-82.1%. Topical steroids were not associated with a statistically significant improvement in clinical symptoms (2.72, 0.9 to 8.23), probably in part because of the high placebo response rates. Topical steroids were well tolerated among patients in all studies. Esophageal candidiasis was consistently noted in the treatment arm, ranging from 4% to 26% of treated patients. However, this responded well to oral nystatin treatment. No significant adrenal suppression was reported.
Although corticosteroids are considered the first line drug treatment for patients diagnosed with eosinophilic esophagitis, currently none are approved for this purpose by the Food and Drug Administration. A phase II multicenter trial using budesonide oral suspension in 93 patients showed that this treatment significantly improved symptomatic, endoscopic, and histologic parameters compared with placebo.175 A multicenter phase III trial is currently under way.
Although most patients respond with histologic improvement of eosinophilia to topical corticosteroids or dietary elimination, a subset of patients will be refractory to standard treatment. Alternative drugs have been explored, including a leukotriene receptor antagonist (montelukast), mast cell stabilizer (cromolyn sodium), and antihistamine, but benefits have been limited.176177 A CRTH2 antagonist (OC000459) showed only modest improvement in esophageal eosinophilia and symptoms in adults with refractory disease.178 Azathioprine and 6-mercaptopurine were effective in case series of patients with steroid dependent eosinophilic esophagitis,179 but potential side effects make these drugs experimental treatments at this time.
Biologic drugs targeting IL-5, IL-13, and IgE have been investigated as potential treatments for refractory eosinophilic esophagitis (table 3⇓). Although antibodies against IL-5, including mepolizumab and reslizumab, have been shown to decrease the esophageal eosinophil count they did not lead to histologic remission or clinical improvement.180181182183184 A recent RCT comparing two dose levels of RPC4046, an IL-13 blocker, and placebo showed that RPC4046 significantly reduced esophageal mean eosinophil count and improved endoscopic features at both doses.185 Several trials have investigated the use of an anti-IgE antibody, omalizumab, but no clinical or histologic disease remission was seen, probably because eosinophilic esophagitis is not primarily an IgE-mediated disease.187188 Further research is needed, but a combination of drugs that inhibit mast cells and eosinophils might be effective for refractory disease.
Eosinophils and mast cells produce TGF-β, which in turn recruits fibroblasts, promotes epithelial-mesenchymal transition, and increases smooth muscle contractility—all of which contribute to esophageal remodeling in eosinophilic esophagitis (fig 2⇑).506063 Ongoing fibrotic changes result in the phenotype of strictures, rings, and narrow caliber esophagus seen in adult patients.2061120189190191 Although some anti-inflammatory agents may improve fibrosis at the microscopic level, in most patients esophageal dilatation is key to providing immediate and long lasting relief of dysphagia.
Case reports and series have described a deep esophageal tear associated with pain, need for hospital admission, and perforation in patients with eosinophilic esophagitis after esophageal dilatation. This resulted in an early consensus recommendation that whenever possible medical or dietary therapy for eosinophilic esophagitis should be attempted before performing esophageal dilatation.8 However, multiple large scale studies have reported low rates (<1%) of severe complications from esophageal dilatation in eosinophilic esophagitis,192193194195196197198199 prompting the inclusion of esophageal dilatation as a potential treatment in symptomatic patients in recent guidelines.1 A systematic review of studies involving esophageal dilatation in eosinophilic esophagitis found an overall perforation rate of 0.6%,200 similar to that of dilatations done for other indications. Esophageal dilatation has been shown to be highly effective although often associated with post-procedural retrosternal pain (74%). However, this was rarely severe and all patients were agreeable to repeating dilatation if needed.193
The latest practice guideline was published in 2013.7 Similar to previous guidelines, eosinophilic esophagitis is defined as peak eosinophil density of ≥15 cells and exclusion of secondary causes of esophageal eosinophilia (box 1). The 2013 guideline emphasizes the exclusion of PPI-REE with a two month course of PPI followed by endoscopy with biopsies. Recommendations on treatment endpoints continue to be “conditional” owing to the lack of supporting data, but symptom assessment was again discouraged as the sole parameter to monitor treatment response. Either topical steroids or dietary elimination is recommended as first line treatment for the management of esophageal eosinophilic inflammation. Both treatments have similar efficacies, so the choice is usually based on patient preference and resources, and esophageal dilatation is recommended in symptomatic patients with evidence of strictures. Owing to the propensity for progressive fibrosis, maintenance treatment is recommended, particularly in those with severe symptoms and a history of high grade strictures, to prevent complications. Figure 5⇓ shows a treatment algorithm based on currently available guidelines. Most recommendations are “conditional” or based on relatively low quality of evidence and will probably change with the emergence of higher quality data.
Box 1: Secondary causes of esophageal eosinophilia
Proton pump inhibitor responsive esophageal eosinophilia
Gastroesophageal reflux disease
Eosinophilic gastrointestinal diseases
Connective tissue disease
Graft versus host disease
In the past two decades, eosinophilic esophagitis has evolved from a relatively unknown and underdiagnosed entity to a well defined disease now known to be a major cause of foregut symptoms in children and adults. Estimated annual healthcare costs are up to $1.4bn (£1.04bn; €1.18bn) in the US.201 In addition, with the increasing recognition of eosinophilic esophagitis as a global health concern, research to expand our understanding of its pathogenesis and treatment has increased in recent years.
In vitro and in vivo studies, animal models, drug trials, and genetic analyses have led to a considerable breakthrough in the understanding of the pathophysiology of eosinophilic esophagitis. We have also made strides in the understanding of the genetic and environmental risk factors that affect the phenotypic expression of the disease. The lack of a clear endpoint in the assessment of disease activity and treatment response has been a limiting factor in clinical trials. Recent advances in patient reported outcome assessment and a standardized method to evaluate endoscopic features will improve the interpretability of future clinical trials. Disease prognostication and risk stratification may be possible with novel technology such as FLIP and mucosal impedance. From a management standpoint, phase III drug trials are under way for eosinophilic esophagitis specific drugs, and effort is being made to optimize dietary elimination protocols to reduce upfront restriction of diet and decrease the number of endoscopies. Novel treatment options, such as biologic drugs, are being investigated for treatment of refractory disease. Many areas of controversy remain in the management of eosinophilic esophagitis, and guidelines will probably be updated in the coming years as a result of fast evolving data on non-invasive diagnostic strategies, new treatment modalities, and the long term prognosis of eosinophilic esophagitis.
Questions for future research
What risk factors (such as a Western diet or dysbiosis) can predict a rapid progressive or fibrostenotic eosinophilic esophagitis phenotype?
Are there any non-invasive biomarkers or testing strategies that could be used to assess disease activity?
Could combination therapy be useful in the treatment of patients with refractory disease?
Could treatments that target barrier function and dysbiosis be useful as primary or adjunctive therapy in eosinophilic esophagitis?
What is the long term prognosis of eosinophilic esophagitis and proton pump inhibitor responsive esophageal eosinophilia? Do maintenance therapies change the clinical outcome?
How patients were involved in the creation of this article
To engage patient participation and contribution to this review, we reviewed the “Frequently asked questions” on the Consortium of Eosinophilic Gastrointestinal Disease Researchers (CEGIR) website (see appendix 2 at end of document) and incorporated answers to these patient oriented questions in our review. In addition, we solicited a list of 10 questions polled from an eosinophilic esophagitis patient support group on social media. These questions were answered and are included in appendix 3. Our review was also critically reviewed by two adult patients with eosinophilic esophagitis before submission.
Series explanation: State of the Art Reviews are commissioned on the basis of their relevance to academics and specialists in the US and internationally. For this reason they are written predominantly by US authors
Contributors: JWC was responsible for the initial outline, literature search, drafting of the initial and revised manuscript; she is also the guarantor. JYK was responsible for the initial concept and outline of the work and offered critical revisions and final approval of the manuscript.
Competing interests: We have read and understood BMJ policy on declaration of interests and declare that we have no competing interests.
Provenance and peer review: Commissioned; externally peer reviewed.