|Home | About | Journals | Submit | Contact Us | Français|
Once considered a rare condition, eosinophilic esophagitis is now one of the most common conditions diagnosed during the assessment of feeding problems in children and during the evaluation of dysphagia and food impaction in adults.1 The entity exists worldwide but has been most extensively studied in Western countries, where its prevalence has been estimated to be 0.4% among all children and adults.2 Whether eosinophilic esophagitis is truly a new disease or simply a recently recognized one is uncertain.3 In this review, we consider the diagnostic criteria, pathophysiological and clinical features, and treatment of this increasingly prevalent disease.
Esophageal eosinophilia was initially considered solely a manifestation of gastroesophageal reflux disease (GERD). However, in the mid-1990s, experienced clinicians identified esophageal eosinophilia in both adults and children who had other symptoms. Neither the clinical symptoms nor the histologic changes in these patients responded to acid suppression and antireflux surgery, which suggested that the condition was distinct from GERD. Two studies of case series4,5 and evidence of the resolution of esophageal eosinophilia in response to therapy with an elemental-formula diet6 suggested that eosinophilic esophagitis was a unique entity. However, clear diagnostic criteria were lacking.
More recently, evaluation and treatment recommendations have been developed on the basis of clinical experiences from different medical subspecialties and the increasing body of knowledge derived from clinical and basic research.7, 8 Eosinophilic esophagitis is currently defined as a chronic, immune-mediated or antigenmediated esophageal disease characterized by symptoms related to esophageal dysfunction and eosinophil-predominant inflammation. The dominant antigens that mediate this disease appear to be food-based. Clinically, eosinophilic esophagitis is defined by several components. First, symptoms include — but are not limited to — feeding problems, vomiting, and abdominal pain in children and dysphagia and food impaction in adolescents and adults. Second, esophageal mucosal eosinophilia of at least 15 eosinophils per high-power field is present. Other causes of these findings, particularly GERD, must be ruled out.7, 8 However, GERD may be difficult to rule out, because neither the response to proton-pump inhibitors nor the duration of exposure to esophageal acid, measured by means of ambulatory pH monitoring, definitively distinguishes GERD from eosinophilic esophagitis.9 Other causes of esophageal eosinophilia (e.g., parasitic infection, allergic vasculitis, esophageal leiomyomatosis, and Crohn's disease of the esophagus) are rare.
The increasing prevalence of eosinophilic esophagitis has focused attention on environmental exposures. Birth by cesarean section, premature delivery, antibiotic exposure during infancy, food allergy, lack of breast-feeding, and living in an area of lower population density have all been associated with eosinophilic esophagitis,10,11 which may imply that altered stimulation of the immune system at an early age confers a predisposition to this disease. Studies have also suggested that a lack of early exposure to microbes11 and an altered microbiome may play a role, as has been suggested for other atopic diseases, such as asthma and atopic dermatitis.12 There is speculation that these exposures early in life create an epigenetic signature that increases the probability of developing eosinophilic esophagitis.
The male predominance of eosinophilic esophagitis, as well as studies of family history and twin concordance and genomewide association studies, suggest that there is a genetic component to eosinophilic esophagitis. In almost every study, male sex has a 3:1 preponderance.13 A family history of eosinophilic esophagitis is frequently reported, with the heritability risk estimated to be 2% on the basis of results from a nuclear family– based cohort of 914 probands with eosinophilic esophagitis and 63 twin probands.14 Genomewide association studies have reported three genes with proposed functional sequelae (the genes encoding thymic stromal lymphopoietin,15 eotaxin-3 [also called chemokine C-C motif ligand 26],16 and calpain-1417, 18) as being altered in eosinophilic esophagitis.
Assessment of esophageal tissues from patients with eosinophilic esophagitis has revealed a striking pattern of dilated interepithelial spaces,19 altered epithelial barrier function,20 and down-regulation of proteins associated with barrier function (filaggrin and zonulin-119) and adhesion molecules (desmoglein-121) (Fig. 1). Interleukin-13 has been shown to down-regulate desmoglein-121 and filaggrin22 in vitro. Altered epithelial permeability23 can lead to a permissive environment that enhances antigen presentation, which in turn leads to recruitment of eosinophils.
Several lines of evidence support the concept that eosinophilic esophagitis is an entity mediated by type 2 helper T (Th2) cell activity and induced primarily by food antigens. Studies of case series have repeatedly reported that patients with eosinophilic esophagitis have environmental and food hypersensitivity, have a response to dietary elimination of food antigens, and have a relapse with reintroduction of similar food antigens.6,24 Whereas food antigens are primarily identified as the allergic triggers, inhaled aeroallergens have also been implicated in some patients; this finding may represent a manifestation of cross-sensitization to food allergens.25 Indeed, murine models of eosinophilic esophagitis can be induced by sensitizing and challenging mice with aeroallergens and ovalbumin26,27 or by overexpressing the cytokines seen in human eosinophilic esophagitis in the animals.28 Several lines of evidence suggest that interleukin-5 and interleukin-13 may play a role in the pathogenesis of eosinophilic esophagitis. Interleukin-5–null mice develop less esophageal eosinophilia in allergen-induced models, as compared with wild-type mice. Interleukin-13 is over-expressed in biopsy specimens from patients with eosinophilic esophagitis and has been found to regulate a number of related cytokines in in vitro studies22 and in studies of mice. In addition, separate prospective trials have shown that treatment with anti–interleukin-5 or anti–interleukin-13 antibodies reduces numbers of esophageal eosinophils in children and adults.29,30 There are also cytokine-directed and thymic stromal lymphopoietin– directed proliferation and recruitment of eosinophils, IgE-bearing mast cells,31 Th2 lymphocytes,32 basophils33 and natural killer cells34 in patients with eosinophilic esophagitis.
The predominant mechanism of food allergy in eosinophilic esophagitis appears to be a non– IgE-mediated process, because omalizumab, an anti-IgE biologic treatment, is not effective in humans,35 esophageal eosinophilia can develop in IgE-null and B-cell–null mice,36 and IgE-based skin testing does not consistently identify food-antigen triggers.37 Food-specific IgG4 was recently shown to be present in the esophageal epithelium and to be reactive to the four most common food-antigen triggers in patients with eosinophilic esophagitis38
An intraluminal balloon device that is used to assess esophageal stricturing and fibrosis by measuring wall compliance may be useful even when stricturing and fibrosis are not detectable endoscopically.39 Early studies with that device showed that patients with a specific distensibility plateau were less likely to have a history of food impaction or to have food impaction during follow-up evaluation. Contrast esophagography in patients with abnormal esophageal-wall compliance often reveals that the esophageal caliber is decreased and the esophagus is diffusely less distensible.40 In vitro exposure of esophageal fibroblasts and smooth-muscle cells to transforming growth factor β (TGF-β) leads to smooth-muscle contraction and fibrosis.41-43
Since the establishment of diagnostic guidelines and a diagnostic code (International Classification of Diseases, Ninth Revision, code 530.13; International Classification of Diseases, Tenth Revision, code K20.0), some studies have estimated the prevalence of eosinophilic esophagitis as between 1 and 5 per 10,000 persons in the United States and Europe, with an increasing prevalence in Asia.44-47 Certain subpopulations, such as patients who are undergoing endoscopy for a history of food impaction, have a markedly higher prevalence, of up to 54%.48
Eosinophilic esophagitis has been described in all age groups, but it predominantly affects white men, with an onset from school age to midlife.46 A personal or family history of atopic disorders, such as asthma, eczema, rhinitis, and anaphylactic food allergy, is common, and these conditions require treatment.49
Children may have a wide variety of nonspecific symptoms, including feeding difficulty, nausea and vomiting, heartburn, and failure to thrive. In contrast, teenagers and adults are more likely to present with dysphagia and episodes of food impaction. Nevertheless, patients of different age groups may have symptoms in common — for example, chronic reflux symptoms. Symptoms may be underestimated as a result of long-standing and subtle accommodation, such as eating slowly, chewing carefully, cutting food into small pieces, lubricating foods with sauces, drinking liquids to dilute foods, and avoiding pills and foods likely to cause dysphagia, such as meats and breads. Patients may be afraid to eat in public places, worrying that they will have difficulty eating. In rare cases, eosinophilic esophagitis may manifest with spontaneous rupture of the esophagus from forceful retching (Boerhaave's syndrome) after a food impaction. Heartburn, especially with the ingestion of alcohol, occurs in 30% of adult patients.13,50 It is unclear whether years of unrecognized childhood subclinical disease or “silent” chronic inflammation precede presentation in adults who have eosinophilic esophagitis or whether adult patients truly have late onset of the disease or a different pathogenesis or phenotype of the disease. In support of the hypothesis that unrecognized childhood subclinical disease or silent chronic inflammation precedes presentation in adults are results from a natural history study in which it was observed that 85% of adults who had untreated symptoms for 20 years presented with esophageal strictures.51
Endoscopy and contrast radiography provide complementary tools for the assessment of the esophagus in eosinophilic esophagitis (Fig. S1 in the Supplementary Appendix, available with the full text of this article at NEJM.org). The most common endoscopic findings are white specks (representative of eosinophilic exudates), mucosal edema, linear furrows, esophageal rings, and strictures.52,53 Chronic remodeling is represented by strictures; a so-called “crepe-paper esophagus,” in which linear tears occur in response to minimal trauma, such as passage of the endoscope50; and the “tug sign,” a firm feeling sensed by the endoscopist when performing an esophageal biopsy.54 A validated endoscopic scoring system was recently developed to standardize the assessment of eosinophilic esophagitis signs that scores the presence of edema, rings, exudates, furrows, and strictures (EREFS).52
The recognition of eosinophilic esophagitis as a diagnostic entity has led to increased use of barium esophagography to evaluate the esophageal lumen in patients with dysphagia. In contrast to the focal distal esophageal strictures found in patients with GERD, strictures in patients with eosinophilic esophagitis may be lengthy and tapered and commonly escape detection during endoscopy but are evident in contrast esophagrams. Two recent studies reported that 71% of adults and 55% of children with eosinophilic esophagitis did not have esophageal narrowing recognized at the time of endoscopy but did have evidence of narrowing at the time of esophagography.55,56
An increased number of eosinophils in the esophageal epithelium, which is a mucosa that is typically devoid of eosinophils, is the histologic hallmark of eosinophilic esophagitis (Fig. 2). A cutoff value of at least 15 eosinophils per high-power field is thought to approach a sensitivity of 100% and specificity of 96% for establishing the histologic diagnosis of eosinophilic esophagitis,57 although patients with lower levels of eosinophilia and phenotypic features of eosinophilic esophagitis have been described.58 Characteristic, but not pathognomonic, features include aggregates of eosinophils or microabscess and eosinophil layering along the luminal surface. Other associated patterns of injury include dilated intercellular spaces, rete-peg elongation, and basal-cell hyperplasia.59 Numbers of inflammatory cells, including lymphocytes, mast cells, and basophils, are also increased in the affected epithelial space.33,60
Complications that may occur in association with eosinophilic esophagitis include esophageal stricture, food impaction, perforation, and malnutrition but not cancer. A number of coexisting conditions have been associated with patients with esophageal eosinophilia, including connectivetissue diseases,61 celiac disease,62 and Crohn's disease.63
Patients with eosinophilic esophagitis may have a mismatch between symptoms and histopathologic features, which creates the need for multiple assessments of disease activity.64 As a result, the short-term treatment goals include the alleviation of symptoms, control of inflammation, and restoration of function. Three approaches can be used to reach these goals: diet, drugs, and dilation.65 When possible, therapy should be provided with a team approach that includes a gastroenterologist, an allergist, and a nutritionist.
In 1995, a study described the favorable effects of an exclusive amino acid–based formula diet (i.e., one devoid of food antigens) in the treatment of 10 children with eosinophilic esophagitis.6 Both the symptoms and the histologic abnormalities improved, but they recurred after resumption of a normal diet. This small study was followed by much larger studies66 and by a similar study involving adults67; in these studies, the patients had a near complete response to therapy, and the stage was set for the routine use of diet therapy in the treatment of eosinophilic esophagitis. The expense and lack of patient interest in an elemental diet led to the development of two other diet strategies (Table 1). The use of skin-prick, atopy-patch, or specific serum IgE testing performed by an allergist to determine targeted diets has been shown to have a high degree of success in children; however, more recent work indicates that this approach may not be as effective as initially thought, with only 45% of patients having a sustained response.68 Nevertheless, elimination diets and the avoidance of food anaphylaxis as directed by an allergist are reasonable treatment options. An alternative diet treatment does not rely on food-allergy testing but rather eliminates the six most commonly identified types of allergenic food (wheat, milk, soy, nuts, eggs, and seafood). This so-called six-food elimination diet was found to improve symptoms and histologic abnormalities in up to 26 of 35 children69 and 32 of 50 adults37 within 6 weeks. Follow-up studies are evaluating a less stringent approach in which fewer foods are initially eliminated.70
Diet therapy can be highly effective and can directly address the underlying allergic mechanism. It may also achieve the goal of identifying a limited number of food antigens that trigger the inflammatory response. For example, because endoscopy with biopsy is currently the only reliable method for the assessment of histologic response, individual patients may undergo multiple endoscopic examinations in an attempt to identify the foods that trigger esophageal eosinophilia.71 Less-invasive testing with a swallowed sponge administered bedside without anesthesia is currently being studied.72 The increased cost of food, poor adherence, and nutritional deficiencies due to the elimination of foods are concerns with diet therapy.
Proton-pump inhibitors may play a role in the diagnostic evaluation of patients with suspected eosinophilic esophagitis and in the care of those with the disease. First, a lack of response to proton-pump inhibitors is currently the only criterion available to rule out GERD as a cause of esophageal eosinophilia. Second, patients with well-established eosinophilic esophagitis can also have symptomatic GERD that is responsive to proton-pump inhibitor treatment and contributes to the development of eosinophilic esophagitis. Third, in vitro studies show that proton-pump inhibitors decrease cytokine secretion from the esophageal epithelium independently of their effect on acid secretion, which leads to the hypothesis that proton-pump inhibitors may provide an antiinflammatory benefit.73 These in vitro studies have further defined a subgroup of patients with an eosinophilic esophagitis phenotype in which the esophageal eosinophilia responded to proton-pump inhibitors, a phenomenon termed proton-pump inhibitor–responsive esophageal eosinophilia. Further study of patients with proton-pump inhibitor–responsive esophageal eosinophilia has shown that they have clinical, histologic, and genetic characteristics similar to those of patients with eosinophilic esophagitis, which supports the possibility that the condition is a subtype of eosinophilic esophagitis.60,74 -76 Nevertheless, the response to proton-pump inhibitors, in contrast to being completely driven by food allergies, may differentiate proton-pump inhibitor–responsive esophageal eosinophilia from eosinophilic esophagitis.
Glucocorticoids target key mechanisms in eosinophilic esophagitis. For example, glucocorticoids decrease fibrosis through the reduction of inflammatory cells.77 Furthermore, increased levels of interleukin-13 (a central regulator of allergic diseases) messenger RNA and the eosinophilic esophagitis transcriptome are largely reversible with glucocorticoid treatment in vivo.78 Although it has not been approved by the Food and Drug Administration, fluticasone administered orally as a spray from a metered-dose inhaler or a viscous preparation of liquid budesonide are the mainstays of pharmacologic therapy for eosinophilic esophagitis.79-84 The efficacy of these topical medications in improving symptoms and histologic abnormalities after 2 to 12 weeks of use ranges from 53%81 to 95%.64,82 Topical glucocorticoids may also reduce the frequency of subsequent food impactions.83 One study has supported the observation that oral viscous budesonide has more prolonged contact with esophageal mucosa and coats a greater length of the esophagus than does nebulized budesonide.84 Recent genomic analyses may help identify whether patients will be more or less likely to have a response to glucocorticoid treatment.79 The potential side effects include local candida infection, adrenal axis suppression, bone demineralization, and diminished growth. Because swallowed topical glucocorticoids undergo first-pass metabolism, such effects appear to be uncommon.85 Nevertheless, the use of systemic glucocorticoids to treat eosinophilic esophagitis leads to similar therapeutic results but incurs a greater risk of side effects than does the use of topical glucocorticoids.80
Esophageal dilation to alleviate esophageal narrowing is a commonly accepted therapy for eosinophilic esophagitis, particularly in older teenagers and adults. Whereas earlier reports suggested a high rate of complications related to dilation in patients with eosinophilic esophagitis, a review of several large series has shown perforation rates of less than 1% (3 of 992 dilations).86 In patients with eosinophilic esophagitis, dilation should be performed gradually over multiple sessions, with an expectation that 75% of patients will have chest pain after the procedure (a course that differs from that in patients with GERD). Dilation successfully addresses the luminal narrowing that can complicate eosinophilic esophagitis but does not treat the underlying inflammatory process.
One of the controversial questions in the management of eosinophilic esophagitis is the role of long-term maintenance therapy. The evidence to date indicates that eosinophilic esophagitis is a not premalignant disease and does not diminish life span. Furthermore, periods of prolonged spontaneous or dilation-induced symptomatic remission may occur without the need for dietary changes or medical therapy.87 However, in most patients, eosinophilic esophagitis is a chronic disease, and if treatment is stopped, inflammation ensues and symptoms recur; quality of life, including vitality and general health scores, is subsequently diminished,88 and complications such as strictures may develop.51 Few studies have examined the long-term benefit of diet or topical glucocorticoid therapy in eosinophilic esophagitis, although both have been proposed as options for maintaining remission.89 The expert consensus opinion is that evidence of chronic remodeling, as in patients with long strictures or a small-caliber esophagus, a history of food impactions or severe symptoms, or rapid recurrence of illness while not receiving therapy, indicates the need for maintenance therapy.7, 8
Since the initial description of eosinophilic esophagitis in case series more than 20 years ago, interest in it has burgeoned, and there is now a greater awareness of the clinical features, basic mechanisms, and effective therapeutic options. The advent of molecular characterizations and diagnostics will lead to the determination of new therapeutic strategies.90 Progress in the treatment of eosinophilic esophagitis will require a better understanding of the implications of this chronic and probably lifelong disease, the development of new therapeutic approaches, and the documentation of the safety and efficacy of long-term therapies.
Figure S1: Endoscopic and Radiologic Findings
Dr. Furuta reports receiving consulting fees from Genentech and grant support from Morphotek and Nutricia, and is co-founder of EnteroTrack. Dr. Katzka reports that his laboratory has provided funding to Capnostics to develop an esophageal sponge. No other potential conflict of interest relevant to this article was reported.
Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.
Glenn T. Furuta, Department of Pediatrics, Section of Gastroenterology, Hepatology, and Nutrition, University of Colorado School of Medicine, and the Gastrointestinal Eosinophilic Diseases Program, Mucosal Inflammation Program, Digestive Health Institute, Children's Hospital Colorado, Aurora.
David A. Katzka, the Division of Gastroenterology and Hepatology, Mayo Clinic, Rochester, MN.