Secondary Logo

Journal Logo

Classification of Gastrointestinal Disease of Infants and Children Due to Adverse Immunologic Reactions Proceedings of a Workshop

Eosinophilic Gastroenteritis

Kelly, Kevin J.

Editor(s): Simpson, Hugh A.; Anderson, John A.

Author Information
Journal of Pediatric Gastroenterology and Nutrition: January 2000 - Volume 30 - Issue 1 - p S28-S35
  • Free


Eosinophilic gastroenteritis refers to a disorder characterized pathologically by the infiltration of eosinophils into the mucosa of the gastrointestinal tract. It has been described in the literature for more than 30 years (1). The associated clinical manifestations take many forms. Clinical symptoms are presumably related to the site in the gastrointestinal tract involved (2). Neither the underlying cause of the eosinophilic infiltration, nor the mechanisms by which the eosinophil infiltrates the gastrointestinal mucosa are yet conclusively known.

The purpose of this article is to present a review of the literature on eosinophilic gastroenteritis to see how well-defined the diagnosis is, and to determine whether it can, in its present form, be included in a standardized diagnostic classification scheme of food-related gastrointestinal hypersensitivity disease syndromes.


Pediatric patients of all ages, infancy (2-4), childhood (5-7), and adolescence (8), have been reported to have eosinophilic gastroenteritis. The disease is also reported in adults, predominantly in the second to the sixth decades of life (9), with reports of occurrence through the seventh decade (10). The disease occurs in both sexes, but males are affected at a slightly higher frequency (11). The disease manifests with an immense variety of chronic and often debilitating gastrointestinal symptoms (Table 1). Many extragastrointestinal clinical manifestations have been reported, but none is consistent (6). Eosinophilic cystitis (12,13), eosinophilic splenitis, and hepatitis (13) have been described. Bone marrow examinations show reactive eosinophilia (6). Analyses of liver biopsy specimens generally reveal eosinophils in the hepatic sinusoids consistent with what might be expected from the peripheral eosinophilia (6,12).

Gastrointestinal symptoms of eosinophilic gastroenteritis


Currently, a few broad and relatively simple parameters are required to make the diagnosis. The diagnosis requires one or more abnormalities of gastrointestinal function and an eosinophilic infiltration of some portion of the gastrointestinal tract. Eosinophilic gastroenteritis is a diagnosis made by a pathologist, with clinical correlation from a practitioner. Peripheral eosinophilia is commonly found, but its presence as a diagnostic criterion is uncertain. Talley et al. (2), reported 40 patients with the diagnosis of eosinophilic gastroenteritis and compared them with 10 control subjects. The data showed that the peripheral eosinophil count was normal in 23% of patients with the diagnosis of eosinophilic gastroenteritis. Furthermore, the 10 control patients in this report with similar symptoms had peripheral eosinophilia but no evidence of eosinophilic infiltration of the gastrointestinal tract mucosa. Additionally, Johnstone and Morson (14) reported in their review of the literature that 24 of 99 patients with eosinophilic gastroenteritis had a normal eosinophil count in the peripheral blood. These findings make the presence of a peripheral eosinophilia an unreliable diagnostic criterion. The absence of parasitic infestation (15) or other extraintestinal diseases associated with eosinophilia are commonly mentioned but are not universally accepted as necessary diagnostic criteria.

Because the diagnostic criteria are so broad and encompassing, a biopsy specimen with nearly any amount of eosinophilic infiltration from a patient with nearly any gastrointestinal symptom raises the possibility of the diagnosis of eosinophilic gastroenteritis. This clinical conundrum is not rare. In our clinical practice, eosinophils are frequently encountered on biopsy of the gastrointestinal tract mucosa (personal observation). If some other disease process is encountered, such as Helicobacter pylori(6), or Crohn's disease (16), the presence of the eosinophil may be interpreted as secondary to that disease. If the eosinophilic infiltration is seen in the distal colon in a breast-fed infant with rectal bleeding, the diagnosis of allergic colitis is entertained. If the eosinophilic infiltration is seen in the distal esophagus in a child with chronic emesis, gastroesophageal reflux disease or allergic eosinophilic esophagitis (17) are considered.

Unfortunately, no standards exist for making the histologic diagnosis (18). No standardized cell-counting system exists, and no standardized system exists to determine when the distribution of the eosinophil within the mucosa constitutes a pathologic process. In a recent publication, Lowichik and Weinberg (19) reported the mucosal eosinophil counts per high-power field in 43 infants and children with no definable gastrointestinal tract symptoms who were medical examiners' cases. The cause of death was varied and included sudden infant death syndrome, trauma, asphyxia, and others. Eosinophils were counted in the stomach, small intestine, colon, and rectum. Eosinophil counts in the gastric fundus and antrum were consistently low. This suggests that when encountered, the finding of an increased density of eosinophils in the gastric mucosa in patients with gastrointestinal symptoms probably represents a true pathologic response. However, in the terminal ileum, cecum, and proximal colon of these asymptomatic patients, eosinophil counts as high as 30 per high-power field were seen. The biopsy specimens collected from these asymptomatic patients showed no significant differences when compared with those collected from children who had been evaluated for symptoms of chronic abdominal pain in the same institution. These observations suggest that increased mucosal eosinophil counts in the lower gastrointestinal tract, even in the presence of chronic gastrointestinal tract symptoms, have no clinical significance (19).

This report further complicates the simple presence of an increased mucosal eosinophil count as a diagnostic criterion by suggesting that differences in eosinophilic densities within different regions of the gastrointestinal tract sampled also occur differently within various regions of the country (19). Finally, different, confusing terminology has been proposed, such as allergic gastroenteropathy (20) or allergic eosinophilic gastroenteropathy. Although these diagnoses also have eosinophilic infiltration detected in mucosal specimen analysis, their diagnostic criteria as yet are poorly defined. Reliable epidemiologic data for eosinophilic gastroenteritis do not exist. Although the disease is sporadic in distribution and familial disease has been reported (21), the prevalence of the diagnosis of eosinophilic gastroenteritis in the general population is unknown.


Can a 77-year-old woman with a 12-year history of intermittent cramping epigastric pain, vomiting, a 30-lb weight loss, ankle-swelling edema, diarrhea, a 30% peripheral eosinophil count, and an eosinophilic infiltration of the small intestinal mucosa (10) have the same disease process as a 25-day-old girl with a 1-week history of nonprojectile, nonbilious vomiting, pylorospasm, and eosinophilic infiltration of the lamina propria and submucosa of the gastric antrum (22)?

Immunologic Findings

The intestinal mucosa has two major and largely opposing functions that must be maintained in a steady state (17). Nutritional uptake and host defenses are shared functions of the surface mucosal cells. These cells are in constant contact with the environment. The mucosal surface is in contact with a wide variety of antigenically active proteins. These include naturally occurring food protein antigens, food antigens that have been altered in meal preparation, or food antigens that have been altered by gastric or duodenal digestive enzyme admixture. Additionally, micro-organisms or swallowed aeroallergens (23) rendered into liquid phase by oropharyngeal secretions present a large antigen load to the intestinal mucosal defense mechanisms. The same enormous surface area through which all absorption and digestion occurs must also provide host defense. The mucous coating, the network of immunocompetent cells, inflammatory cells, nerves, and blood vessels in the lamina propria and submucosa constitute the complex defense system of the intestinal mucosa (16). Monocytes, eosinophils, and tissue mast cells exist in the intestinal tissue and mediate intestinal inflammatory responses. Other cells involved in the inflammatory response, such as the neutrophils and basophils, migrate from the circulation to the sites of intestinal inflammation (16). The constant juxtaposition of these active, immunologic components, capable of mounting a hypersensitivity response, and the large amount of foreign protein antigens predisposes the intestinal mucosa to hypersensitivity reactions (16).

The presumption that the eosinophilic infiltration of the mucosa of the gastrointestinal tract is associated with an adverse, immunologic response has been a well-received hypothesis for many years (3). This association between allergy and the eosinophilic infiltration of gastrointestinal tract mucosa is supported by the frequent coexistence of other atopic disorders, such as asthma (24), seasonal allergic rhinitis, eczema, drug allergies (13), or multiple food allergies (25). The eosinophilic neutrophil has long been viewed as a marker of immune-mediated responses in the mucosa of these other human systems. Eosinophils have been found in bronchoalveolar lavage samples from patients with extrinsic asthma, in mucosal specimens of patients with seasonal allergic rhinitis, and in allergic conjunctivitis. The presence of the eosinophil in mucosal specimens of the gastrointestinal tract raises the same suspicion.

Patients with these associated diagnoses frequently also demonstrate positive food-specific antibodies of the immunoglobulin (Ig)E isotype, detected in vivo by prick skin testing or in vitro by radioallergosorbent testing (RAST) (15). Some researchers report a strong association between the ingestion of foods identified by these methods and the development of gastrointestinal tract symptoms (14,26,27). Others report that antibodies detected in these patients do not successfully identify foods that cause gastrointestinal tract symptoms when ingested (24). Overall, antibody testing is only slightly better than random chance (50%) at identifying specific foods related clinically to the symptomatic manifestations in these patients.

Newer evidence regarding eosinophilic infiltration of the gastrointestinal tract mucosa suggests that the eosinophil may participate in a bidirectional link between the immune system and inflammatory effector cells (16). Interleukin (IL)-5 is a cytokine secreted from the lymphocyte that attracts and activates eosinophils. It is responsible for tissue eosinophilia (28). Once activated, normal human eosinophils themselves are primed to secrete a variety of inflammatory mediators under the influence of IL-3 and IL-5 (29). The eosinophil can also produce the cytokines transforming growth factor, granulocyte-macrophage colony-stimulating factor (GM-CSF), IL-3, and IL-5 (30). These cytokines are known to be chemotactic for the eosinophil (28). They are secreted by eosinophils, partly in an autocrine manner (16). The eosinophil itself can therefore contribute to the propagation of the eosinophilic inflammatory process. Justinich et al. (31), found that infants and children with esophagitis exhibited a higher expression of IL-5 mRNA and protein than did control subjects, with IL-5 localized to the eosinophil.

More recently, Desreumaux et al. (32) compared the mucosal specimens of 10 patients with eosinophilic gastroenteritis with those of 10 control subjects. The number of lymphocytes and eosinophils in the duodenal and colonic mucosa was significantly higher in the patients with eosinophilic gastroenteritis than in the control subjects. There were no differences in the number of neutrophils between the two groups. The results of immunohistochemical staining between these two groups demonstrated the presence of IL-3, GM-CSF, and IL-5 in 9 of 10 patients with an eosinophilic infiltration and none in patients with histologically normal specimens. Ultrastructurally, these cytokines were localized in the granule matrix of the eosinophils. These findings further support the hypothesis that, once activated, eosinophils possess the capacity to synthesize cytokines and regulate their own proliferation and differentiation (32). The capability of expressing IL-5 and GM-CSF has also been demonstrated in eosinophils derived from patients with asthma (28,32). It is unknown whether these findings suggest intrinsic differences in the eosinophil in patients with these diseases, or a common pathogenesis.


Kaijser originally described this disorder in 1937 (1). Subsequently, Ureles et al. (33) were the first to categorize some of the patterns of clinical presentation. They proposed two provisional classifications based on the pathologic and clinical appearances. These patterns of distribution were defined principally by operative findings. They called class I diffuse eosinophilic gastroenteritis and class II circumscribed eosinophilic-infiltrated granuloma (33).

Class I was subdivided by the patterns of disease distribution into polyenteric, monoenteric and regional forms. Involvement of the pyloric region was found in all patients in all groups. The polyenteric type, group A, involved the pylorus and proceeded in an antegrade fashion to involve the small intestine. Microscopically, eosinophilic involvement extended from the submucosa, through the muscle bundles of the muscularis, through the deep submucosa to the serosa. The mucosa in this type was notably free of involvement. Patients with the monoenteric type, or group B, had the same gross and microscopic involvement, but findings in them proceeded retrograde from the pylorus to involve the entire stomach without involvement of the small intestine. Finally, in the group C, or regional, type of diffuse eosinophilic gastroenteritis, the process was more confined. The disease in this category was limited only to the prepylorus and pylorus. This pattern grossly had ill-defined borders, and microscopically, the disease process was more acute. Clinical features of class I patients showed significant gastrointestinal symptoms of acute onset or occurring chronically for as long as 25 years. Surgical exploration was undertaken because of the severity of the unremitting symptoms. Some form of allergic response was suspected in only approximately half the class I patients. No formal allergy evaluations were conducted. Suspicion of allergy in these patients was based only on the presence of other atopic diseases.

Class II, or circumscribed eosinophilic infiltrated granuloma, was subdivided into group A, regional type, and group B, polypoid type. Grossly, both of these groups had characteristic circumscribed pseudotumors. In group A, the lesions were found throughout the gastrointestinal tract; in group B, the lesion was confined to the pyloric antrum. Regardless of location, these polypoid lesions contained granulomas and abundant, mature eosinophils. Group A patients had initial symptoms of cramping pain and a radiographic appearance suggestive of malignancy. Group B patients had ulcer-like symptoms and gastric bleeding. No patients in these two groups had peripheral eosinophilia, and none had signs, symptoms, or history of allergy. Corticosteroids were not typically used in any patients in this series. In one patient with class I disease that relapsed after surgery, steroids were effective. The histologic appearance of other regions of the gastrointestinal tract in these patients was not reported.

Klein et al. (27) added seven new patients to the literature and conducted a review of the literature up to the time of their publication in 1970. They defined three patterns of disease manifestation based on the initial symptoms. They related the clinical manifestations to the area of maximal gastrointestinal involvement and the depth of the disease process (27). The three main patterns were predominant mucosal layer disease, predominant muscle layer disease, and predominant subserosal disease. Patients with the mucosal predominant disease, had hypoproteinemia secondary to protein-losing enteropathy, iron-deficiency anemia from chronic gastrointestinal blood loss, and weight loss from chronic malabsorption. In children, mucosal predominant disease can result in significant growth delay (34). Patients with clinical evidence of predominant muscle layer disease had obstructive symptoms. The principal location for this type was the pyloric region, although involvement of the small bowel was also reported. The cardinal manifestation of the predominant subserosal disease pattern was ascites. This ascitic fluid contained a large number of eosinophils (27). Klein et al. suggested that this disease could be classified into patterns distinctive enough from each other clinically to provide a convenient means of categorization of patients for further study. This orderly classification scheme was accepted and has been referenced frequently in subsequent publications.

That same year, Leinbach and Rubin (35) investigated more thoroughly the relationship between eosinophilic gastroenteritis and the diet. For 2.5 years, they observed a 23-year-old man with eosinophilic gastroenteritis who went through sequential dietary eliminations. Each successive elimination produced temporary improvement of symptoms. Eventually, the patient could only tolerate a soy bean hydrolysate formula. He eventually experienced gastrointestinal symptoms even with this restricted diet. The conclusion in this report was that this disorder was not a simple, reversible allergic reaction to specific foods. They speculated that this disease was a self-perpetuating process symptomatically aggravated by different foods (35). Klein et al. also reported that prick skin testing, IgE levels, and allergen-specific antibodies for evidence of immediate hypersensitivity reaction to food substances was unreliable in eosinophilic gastroenteritis (27). Lastly, in this report, after a food challenge that reproduced clinical symptoms, tissue eosinophilia could not be detected in jejunal specimens. This observation challenged the widely held belief that the presence of the eosinophil was responsible for the observed clinical symptoms.

Cello (15) also reported on the relationship between this diagnosis and the diet. He supported the use of an elimination diet, especially in those patients with an atopic history. He recommended rigid restriction of substances that were known to exacerbate symptoms or were suspected of exacerbating symptoms. He observed and reported that small amounts of foods in sensitive patients could elicit severe abdominal pain. He speculated that the small amounts of foods may also increase eosinophilic mucosal infiltration, although this was not demonstrated or referenced. He also summarized the existing theories of that era, for the mechanisms responsible for the chemotaxis of eosinophils into the gut mucosa: 1) An Arthus-type immediate hypersensitivity may be responsible for attracting eosinophils to the site of antigen-antibody complexes; 2) activation of the standard complement pathway might draw eosinophils to the site of complex deposition by the C-3 receptor on the surface of the eosinophil; and 3) sensitized T cells activated by their respective antigens may release lymphokines capable of attracting peripheral blood eosinophils.

Cello also advanced a hypothetical scheme involving food antigens reacting in the gut wall with food-specific IgE antibodies bound to tissue mast cells at Fc receptor sites. The degranulation of mast cells caused the release of histamine and eosinophilic chemotactic factor of anaphylaxis, causing damage to the mucosal cells. Finally, he speculated that regardless of the mechanism responsible for the chemotaxis of eosinophils to the gut, their major role was to maintain homeostasis and to modulate inflammation. Histaminase, kininase, and phospholipase D were thought capable of neutralizing the effects of inflammatory substances released by the degranulating mast cell (15).

In 1984, Oyaizu et al. (36) presented further evidence for the hypothetical IgE-induced, mast cell-mediated mechanism of eosinophilic chemotaxis in patients with eosinophilic gastroenteritis. They reported IgE-producing plasma cells in the vicinity of the eosinophilic infiltrates. In addition, the surface IgE-positive mast cells in the periphery of the eosinophilic lesions appeared to have their intercellular granules intact, whereas the mast cells within the eosinophilic lesion appeared degranulated.

In 1990, Talley et al. (2) reported 40 patients with eosinophilic gastroenteritis, one of the largest reported clinical experiences with eosinophilic gastroenteritis. They also included, for comparison, 10 patients with similar symptoms but no gastrointestinal tract eosinophilia. They categorized the 40 patients with eosinophilic gastroenteritis according to the classification established by Klein et al. (27) It was the first report, in the 20 years since the Klein classification was proposed, of a study in which patients were comparatively analyzed according to these groupings. Their purpose was to determine whether differences in the clinical and laboratory parameters in fact existed among these groups. The variables analyzed were clinical and laboratory parameters. Groups were compared for symptoms of abdominal pain, nausea, vomiting, weight loss, bloating, and diarrhea. Also, the number of years of symptoms before diagnosis, history of allergy, family history of allergy, food intolerance or reported allergy, the erythrocyte sedimentation rate at initial examination, the blood eosinophil count, the presence of malabsorption or steatorrhea, and the presence of protein-losing enteropathy. For all the variables studied, only bloating as an initial symptom and the presence of a very high eosinophil count demonstrated statistically significant differences among these three groups. Both these differences were associated with the predominant subserosal disease. The nearly complete inability to distinguish any other significant differences among these groups suggests either a high degree of overlap or that this classification scheme may be more artificial than real. Lastly, in a manner similar to that of Leinbach and Rubin (35), Tally et al. (2) further challenged the pathogenic role of the eosinophil. Patients with eosinophilic gastroenteritis could not be distinguished from the control group of patients with no evidence of eosinophilic infiltration. No significant differences in clinical appearance in any of the historical data or laboratory variables studied could be found between patients with eosinophilic infiltration observed in biopsy specimens and the control group of patients with no evidence of eosinophilic infiltration (18). This finding suggests the presence of the eosinophil may not distinguish a distinctive disease state.

In a review article, Min and Metcalfe (21) refute the theory that the eosinophil downregulates immediate hypersensitivity reactions. They reported that eosinophilic cytoplasmic granules induce tissue damage by release of the cytotoxic substances, major basic protein, eosinophilic cationic protein, and eosinophilic-derived neurotoxin (EDN). They further discuss the more recent theories of regulation of eosinophil production and maturation, including the possible role of the T-lymphocyte-derived interleukin, IL-5. In the diagnosis section, they state that the number of eosinophils may increase further if a biopsy site is challenged first with specific food antigens. This is in keeping with the report of Makenda et al. (37) but is contrary to the report of Leinbach and Rubin (35). In their treatment section, Min and Metcalfe (21) suggest a role for the use of an elemental, amino acid-based diet in children, followed by a milk protein hydrolysate. Their recommendation in adult patients is to try an elimination diet in those patients with an atopic history. Finally, they report that removal of selected foods should be based on the results of prick skin tests and RASTs. This is also contrary to the report of Leinbach and Rubin (35). Also, Min and Metcalfe state that pediatric patients with eosinophilic gastroenteritis usually respond to an elimination diet when the disease is related to food sensitivity and that frequently the sensitivity to foods resolves spontaneously by the age of 2 or 3 years. These two statements were not referenced.

Bischoff (16) presents a comprehensive review of the current understanding of mucosal allergy. He discusses in depth the role of mast cells and the eosinophil granulocyte in the gut. He further supports the theory that the eosinophil participates in and propagates the inflammatory process. Once activated, the eosinophil releases proinflammatory mediators such as cationic proteins, leukotrienes, prostaglandins, and platelet-activating factor. Eosinophils are thought also to exert cytotoxic effects by producing oxygen free radicals and peroxidase. In contrast, he supports the suggestion that the mast cell may be involved in the process of tissue repair. He based these statements on the appearance of mast cells at the sites of wound healing and bone fractures. In addition, mast cells were present in increased numbers at the line of demarcation between actively inflamed and normal intestinal mucosa in patients with active ulcerative colitis. Lastly, mast cell accumulation is also reported in fibrotic tissues such as neurofibroma and liver fibrosis, suggesting a role for mast cells in the induction of fibrotic tissue transformation.

Bischoff also describes well some of the challenges faced when dealing with symptoms of gastrointestinal food allergy. If an immune-mediated cause is suspected, he first recommends exclusion of all the nonimmunologic causes of adverse reactions to foods, such as lactose intolerance. His evaluation then includes standard immune testing and a colonoscopic allergen provocation (COLAP) test (16). This is similar to established provocation techniques for the nasal, conjunctival, and bronchial mucosa. In an unpublished observation, he reports that antigen-induced wheal-and-flare reactions correlate with intestinal mast cell and eosinophil activation and with the patients' histories of adverse reactions to food. He endorses the double-blind, placebo-controlled food challenge as a means of confirmation of suspected food antigens. Finally, he speculates that the detection of mediators derived from eosinophils and mast cells in intestinal lavage fluids or stool samples may be helpful in initial diagnosis and, possibly, in prospective management of disease activity (16).


Since the provisional classification scheme was first proposed by Ureles et al. (33) in 1961, the literature has increased an average of only 10 articles per year on the subject of eosinophilic gastroenteritis. Most of these publications enrich the literature with presentations of unusual symptoms (22,24,37-41) or unusual complications (6) of this disease or novel methods of establishing the diagnosis (41). New therapeutic interventions have also been reported (25,42,43). Studies advancing the understanding of the mechanism of this disorder are rare, and no studies have been able to interrelate the laboratory findings, the radiographic findings, or the clinical findings and the initial signs and symptoms to new theories of the proposed mechanisms involved. Why is this?

The broad diagnostic criteria of eosinophilic gastroenteritis imposes a clinical quandary. Will all patients who have met the criteria and have been given this diagnosis be shown to have the same pathologic process? This question will probably be answered as the knowledge base expands, the ability to understand the physiologic mechanisms that influence the eosinophil increases, and we develop the ability to describe these patients reproducibly in a uniform fashion among centers. The potential application of new information to patients who are given the diagnosis of eosinophilic gastroenteritis is limited by the difficulties associated with the performance of prospective clinical trials. Principal among these difficulties is the attainment of control subjects for prospective comparison studies. Two recent reports (2,32) have compared patients with the diagnosis of eosinophilic gastroenteritis with control subjects. As previously discussed, the control group used by Talley et al. (2) comprised patients with similar symptoms but no eosinophilic infiltration detected in mucosal specimens. Aside from the variability of the presence of the peripheral eosinophilia among these groups as previously discussed also in this report, five patients with the diagnosis of eosinophilic gastroenteritis did not show evidence of increased eosinophilic infiltration in results of a follow-up mucosal biopsy. This demonstrates the poor reproducibility of the findings of this diagnosis and raises concern regarding the suitability of the control population.

Desreumaux et al. (32) compared the immunoreactivity for three cytokines within the duodenal and colonic mucosa from 10 patients with eosinophilic gastroenteritis with those of 10 sex- and age-matched control patients with irritable bowel syndrome. They found significant increases in the number of eosinophils in the biopsy specimen and in the three cytokines studied in the patients with the eosinophilic infiltration. No effort was made in this study to link immunohistochemical findings with the initial clinical symptoms or with the location of the eosinophilic infiltration within the gastrointestinal tract. Patients with eosinophilic gastroenteritis had symptoms of abdominal pain, nausea, diarrhea, vomiting, and ascites. All 10 patients had both upper endoscopy and colonoscopy performed. Because these investigators segregated their patients solely according to the presence of eosinophils in mucosal specimens, these findings have advanced the knowledge of the mechanisms behind the recruitment of eosinophils to the gastrointestinal tract mucosa. The study design did not improve the knowledge of the relationship of the eosinophil to the clinical symptoms. Because irritable bowel syndrome also has symptoms of abdominal pain, nausea, diarrhea, and vomiting (44), the presence of the eosinophil in the mucosa of the patients with eosinophilic gastroenteritis in Desreumaux et al. cannot be confidently related to the symptoms. If the presence of the cell is, in fact, related to the symptomatology, patients with eosinophils should be distinguishable from patients without eosinophils.

Whether the eosinophil is central to the symptomatic manifestations or is a marker of another pathologic process that is responsible for the symptoms remains unknown. Does our difficulty with the identification of an appropriate control population arise from the possibility that the disease referred to as eosinophilic gastroenteritis may in fact be too broad and too encompassing? Is eosinophilic gastroenteritis a unifying diagnosis or a group of heterogeneous disorders unified only in the histologic appearance of the eosinophil? Is eosinophilic gastroenteritis specifically a food-related gastrointestinal hypersensitivity disease syndrome, or could the eosinophilic response be a local immune-mediated response to other foreign antigens present in the gastrointestinal tract mucosa?

The eosinophil is found in diseases of other organ systems, such as the lung, conjunctiva, and skin. The presence of the eosinophil in these other systems influences clinicians to suspect allergic disease. If the symptoms improve with the removal of an extrinsic protein antigen identified from the local environment, the suspicion of allergy is supported. If the symptom in these other systems cannot be improved by the removal of a commonly encountered environmental protein antigen, symptoms typically can be improved by the use of antiinflammatory medications while the search for the offending antigen is broadened. During the evaluation of a child with audible wheezing and skin rash, a bronchoalveolar lavage and skin biopsy may demonstrate the presence of an eosinophilic tissue response. That child may receive a diagnosis of asthma and atopic dermatitis and an exhaustive search for the environmental antigen or antigens associated with the symptoms and the eosinophilic response would be undertaken. Would these patients be given the diagnosis of "eosinophilic asthma" or "eosinophilic eczema," and would they be treated with long-term systemic corticosteroids without the benefit of a comprehensive search for the antigen responsible for triggering the wheezing and rash?

In diseases involving an eosinophilic infiltration in these other systems, the simple presence of the cell does not make a definitive diagnosis. It raises a clinical suspicion. In contrast, however, in the gastrointestinal system of patients with clinical symptoms, the simple presence of this cell provides a definitive diagnosis. Blackshaw (18) a pathologist from the United Kingdom, and Levison published an in-depth review article in 1986 in which they discussed inflammatory fibroid polyps and eosinophilic gastroenteritis. In their introduction, they stated the following: "Eosinophilic infiltration, localized or diffuse, of the gastrointestinal tract cannot be regarded as a definitive diagnosis; neither can a classification of disease, in which this is a feature, be founded on the degree or site of such infiltration. Eosinophilia is merely a manifestation of inflammation, often with an allergic basis, and accompanies a number of well defined clinicopathological conditions, some of which have a known cause-for example, parasitic infestation-but most of which do not."

Eosinophils are normally present in the lamina propria of the gut mucosa, and there is no generally agreed upon number that is considered pathologic. Groups of eosinophils are not normally found in the submucosa, muscle coat, or serosa. The common diagnostic dilemma, however, is not whether there are increased numbers of eosinophils present in the lesion, but rather, the pathologic importance of the eosinophilic infiltrate in the lesion. An eosinophilic infiltrate in the bowel may either point to the nature of the underlying condition or may obscure it, and the confusing studies on the subject reflect the diagnostic, conceptual, and terminological difficulties entailed (18).

The questions and the concepts raised by these statements suggest an overemphasis on the presence of the eosinophil. Can these probing questions be answered in clinical practice? Can we answer these questions for this disease as it exists in its present form?

Our collective diagnostic acumen regarding patients with eosinophilic mucosal infiltration has improved greatly. As little as 10 years ago, patients with diagnoses such as allergic colitis or eosinophilic esophagitis associated with gastroesophageal reflux or allergy would have been classified as having eosinophilic gastroenteritis because their symptoms fulfilled the broad criteria for the diagnosis. They are now, however, classified, approached, and treated individually. As our knowledge of gastrointestinal illness continues to expand, the likelihood is great that progressively more diagnoses involving the mucosal eosinophil will be removed from the broad diagnostic category of eosinophilic gastroenteritis. Talley has shown that the preliminary classification scheme proposed by Klein (27) is not a reliable way to distinguish differences between patients with eosinophilic infiltration. The overlap in this classification scheme is great, and it does not easily allow segregation of patients into groupings amenable to prospective study. Collectively, the participants in the International Life Sciences Institute symposium believed that patients with eosinophilic infiltration of their gastrointestinal tract mucosa could best be incorporated into a standardized diagnostic classification scheme of food-related gastrointestinal hypersensitivity disease syndromes if patients were separated into categories by symptoms and region of the gastrointestinal tract involved.

The new classification system proposed for food-related gastrointestinal hypersensitivity disease syndromes attempts to classify diseases by age, symptoms, and the region of the gastrointestinal tract most likely involved. Segregation of patients for prospective analysis by region of involvement is important because of the likelihood that this diagnosis involves an interaction with a variety of possible antigens within the intestinal lumen. Protein antigens with potential immunologic activity vary widely within different regions of the gastrointestinal tract. Furthermore, as ingested protein antigens course through the gastrointestinal tract and are subjected to disruptive mechanical forces and digestive enzyme admixture, their antigenic properties are likely to change. These include, but are not limited to, food proteins, microbial organisms, and swallowed environmental airborne proteins. If this classification scheme is accepted and uniformly applied, it is hoped that our collective ability to perform prospective clinical research will improve.

We should focus our investigative efforts away from the eosinophil as a central diagnostic finding. We should also focus our investigations of the immunologic mechanisms of chronic gastrointestinal symptoms more broadly to include patients with any gastrointestinal tract symptoms that cannot be attributed to known gastrointestinal diseases, regardless of age or symptoms. To restrict immunologic investigations in gastroenterology only to patients with eosinophilic infiltration limits our ability to identify immunologic influences that may have an association with other common symptoms. These investigative efforts should occur, therefore, independent of the presence of the eosinophil. By organizing our approach in this way, we may increase our ability of achieving a greater understanding of the immunologic responses of each region in the gastrointestinal tract to its local environment. As this knowledge increases, our ability to resolve the clinical quandary presented by this complicated disorder is also likely to improve.


1. Kaijser R. Kenntnis der Allergischen Affektionen des verdauungskanals vom Standpunkt des Chirurgan aus. Arch Klin Chir 1937;188:36-64.
2. Talley NJ, Shorter RG, Phillips SF, Zinsmeister AR. Eosinophilic gastroenteritis: A clinicopathological study of patients with disease of the mucosa, muscle layer, and subserosal tissues. Gut 1990;31:54-8.
3. Kay MH, Wyllie R, Steffen RM. The endoscopic appearance of eosinophilic gastroenteritis in infancy. Am J Gastroenterol 1995;90:1361-2.
4. Suen KC, Burton JD. The spectrum of eosinophilic infiltration of the gastrointestinal tract and its relationship to other disorders of angiitis and granulomatosis. Hum Pathol 1979;10:31-43.
5. Whitington PF, Whitington GL. Eosinophilic gastroenteropathy in childhood. J Pediatr Gastroenterol Nutr 1988;7:379-85.
6. Deslandres C, Russo P, Gould P, Hardy P. Perforated duodenal ulcer in a pediatric patient with eosinophilic gastroenteritis. Can J Gastroenterol 1997;11:208-12.
7. Caldwell JH, Mekhjian HS, Hurtubis PE, Beman FM. Eosinophilic gastroenteritis with obstruction - immunological studies of seven patients. Gastroenterology 1978;74:825-8.
8. Moore D, Lichtman S, Lentz J, Stringer D, Sherman P. Eosinophilic gastroenteritis presenting in an adolescent with isolated colonic involvement. Gut 1986;27:1219-22.
9. Edelman MJ, March TL. Eosinophilic gastroenteritis. Am J Roentgenol Radium Ther Nucl Med 1964;91:773-8.
10. Kaplan SM, Goldstein F, Kowlesar OD. Eosinophilia gastroenteritis-report of a case with malabsorption and protein-losing enteropathy. Gastroenterology 1970;58:540-5.
11. Jacobson LB. Diffuse eosinophilic gastroenteritis: An adult form of allergic gastroenteropathy. Am J Gastroenterol 1978;54:580-8.
12. Gregg JA, Utz DC. Eosinophilic cystitis associated with eosinophilic gastroenteritis. Mayo Clin Proc 1974;49:185-7.
13. Robert F, Omura E, Durant JR. Mucosal eosinophilic gastroenteritis with systemic involvement. Am J Med 1977;62:139-43.
14. Johnstone J, Morson B. Eosinophilic gastroenteritis. Histopathology 1978;2:335-48.
15. Cello JP. Eosinophilic gastroenteritis: A complex disease entity. Am J Med 1979;67:1097-104.
16. Bischoff SC. Mucosal allergy: Role of mast cells and eosinophil granulocytes in the gut. Baillieres Clin Gastroenterol 1996;10:443-59.
17. Kelly KJ, Lazenby AJ, Rowe PC, Yardley JH, Perman JA, Sampson HA. Eosinophilic esophagitis attributed to gastroesophageal reflux: Improvement with an amino-acid based formula. Gastroenterology 1995;109:1503-12.
18. Blackshaw AJ, Levison DA. Eosinophilic infiltrates of the gastrointestinal tract. J Clin Pathol 1986;39:1-7.
19. Lowichik A, Weinberg AG. A quantitative evaluation of mucosal eosinophils in the pediatric gastrointestinal tract. Mod Pathol 1996;9:110-4.
20. Moon A, Kleinman RE. Allergic gastroenteropathy in children. Ann Allergy Asthma Immunol 1995;74:5-12.
21. Min KU, Metcalfe DD. Eosinophilic gastroenteritis. Immunol Allergy Clin North Am 1991;11:799-813.
22. Blankenberg FG, Parker BR, Siblery E, Kerner JA. Evolving asymmetric hypertrophic pyloric stenosis associated with histologic evidence of eosinophilic gastroenteritis. Pediatr Radiol 1995;25:310-11.
23. Puente S, Iniguez A, Subirats MJ, Alonso MJ, Polo F, Moneo I. Gastroenteritis eosinofilica causada por sensibilizacion a polen de abeja. Med Clin (Barc) 1997;108:698-700.
24. Park HS, Kim HS, Jang HJ. Eosinophilic gastroenteritis associated with food allergy and bronchial asthma. J Korean Med Sci 1995;10:216-9.
25. Justinich C, Katz A, Gurbindo C, et al. Elemental diet improves steroid-dependent eosinophilic gastroenteritis and reverses growth failure. J Pediatr Gastroenterol Nutr 1996;23:81-5.
26. Katz A, Goldman H, Grand R. Gastric mucosal biopsy in eosinophilic (allergic) gastroenteritis. Gastroenterology 1977;73:705-9.
27. Klein NC, Hargrove RL, Sleisenger MH, Jeffries GH. Eosinophilic gastroenteritis. Medicine 1970;40:299-319.
28. Kelso A. Cytokines: Structure, function and synthesis. Curr Opin Immunol 1989;2:215-25.
29. Takafuji S, Bischoff S, DeWeck A, Dahinden C. IL-3 and IL-5 prime normal human eosinophils to produce leukotriene C4 in response to soluble agonists. J Immunol 1991;147:3855-61.
30. Weller P. The immunobiology of eosinophils. N Engl J Med 1991;324:1110-8.
31. Justinich C, Kalafus D, Fitzgerald J, et al. Increased expression of interleukin-5 in reflux esophagitis in children (abstract). Gastroenterology 1995;108:126A.
32. Desreumaux P, Bloget F, Seguy D, et al. Interleukin 3, granulocyte-macrophage colony- stimulating factor, and interleukin 5 in eosinophilic gastroenteritis. Gastroenterology 1996;110:768-74.
33. Ureles AL, Alschibaja T, Lodico D, Stabins SJ. Idiopathic eosinophilic infiltration of the gastrointestinal tract, diffuse and circumscribed. Am J Med 1961;30:899-909.
34. Caldwell JH, Sharma HM, Hurtubuse PE, et al. Eosinophilic gastroenteritis in extreme allergy: Immunopathological comparison with non-allergic gastrointestinal disease. Gastroenterology 1979;77:560-4.
35. Leinbach GE, Rubin CE. Eosinophilic gastroenteritis: A simple reaction to food allergens. Gastroenterology 1970;59:874-89.
36. Oyaizu N, Uemura Y, Izumi H, Sotokichi M, Nishi M, Hioki K. Eosinophilic gastroenteritis, immunohistochemical evidence for IgE mast cell mediated allergy. Acta Pathol Jpn 1985;35:759-66.
37. Makenda C, Phillips AD, Briddon A, Walker-Smith JA. Quantitative analysis of small intestinal mucosa in cow's milk sensitive enteropathy. J Pediatr Gastroenterol Nutr 1984;3:349-56.
38. Mahajan L, Wyllie R, Petras R, Steffen R, Kay M. Idiopathic eosinophilic esophagitis with stricture formation in a patient with longstanding eosinophilic gastroenteritis. Gastrointestinal Endosc 1997;46:557-60.
39. Kristopaitis T, Neghme C, Yong SL, Chejfec G, Aranha G, Keshavarzian A. Giant antral ulcer: A rare presentation of eosinophilic gastroenteritis: Case report and review of the literature. Am J Gastroenterol 1997;92:1205-8.
40. Matsushita M, Hajiro K, Morita Y, Takakuwa H, Suzaki T. eosinophilic gastroenteritis involving the entire digestive tract. Am J Gastroenterol 1995;90:1868-70.
41. Vara-Thorbeck C, Toscano-Mendez RJ, Osorio D. Eosinophilic gastroenteritis: Diagnostic laparoscopy. Surg Laparosc Endosc 1997;7:66-9.
42. Perez-Millan A, Martin-Lorente JL, Lopez-Morante A, Yuguero L, Saez-Royuela F. Subserosal eosinophilic gastroenteritis treated efficaciously with sodium cromoglycate. Dig Dis Sci 1997;42:342-4.
43. DiGioacchino M, Pizzicannella G, Fini N, et al. Sodium cromoglycate in the treatment of eosinophilic gastroenteritis. Allergy 1990;45:161-66.
44. Sleisenger MH, Fordtran JS. Gastrointestinal disease pathophysiology, diagnosis, management. Philadelphia, PA: Saunders, 1989:1402-18.

Section Description

Washington, D.C., November 16-17, 1998

Sponsored by the International Life Sciences Institute (ILSI) Allergy and Immunology Institute, and cosponsored by the American Academy of Allergy, Asthma & Immunology, the Jaffe Family Foundation and Elliot Roslyn Jaffe Food Allergy Institute, and the ILSI branches in Argentina, India, Japan, and Mexico.

Publisher's Note

The opinions expressed in this presentation are those of the authors and are not attributable to the workshop sponsors or the publisher, editors, or editorial board of Journal of Pediatric Gastroenterology and Nutrition, European Society of Paediatric Gastroenterology, Hepatology and Nutrition, or the North American Society for Pediatric Gastroenterology and Nutrition. Clinical judgment must guide each physician in weighing the benefits of treatment against the risk of toxicity. References made in the articles may indicate uses of drugs at dosages, for periods of time, and in combinations not included in the current prescribing information.

© 2000 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,