Secondary Logo

Journal Logo

Case Reports

Eosinophilic Gastroenteritis Presenting as Acute Gastric Perforation

Siaw, Emmanuel KO*; Sayed, Kadria; Jackson, Richard J

Author Information
Journal of Pediatric Gastroenterology and Nutrition: November 2006 - Volume 43 - Issue 5 - p 691-694
doi: 10.1097/01.mpg.0000239996.66011.89
  • Free

Eosinophilic gastroenteritis (EG) is a chronic recurrent disorder with an unknown etiology. This rare condition is partly characterized by abdominal pain, abdominal distension, fever, vomiting and diarrhea—symptoms that are not specific to this condition and are also characteristic of several other gastrointestinal (GI) disorders. Leinbach and Rubin (1) proposed 3 criteria to diagnose EG, one of which was for the patient to exhibit eosinophilic leukocyte infiltrates in the GI tract.

Food allergies (1), parasitic infestation, drug reactions, Crohn disease and hypertrophic pyloric stenosis should be considered in the differential diagnoses (2). Leinbach and Rubin (1), studying patients with EG more than 30 years ago, suggested that food allergies did not cause the disorder. Caldwell et al. (3) reported a pediatric case of EG in an allergic boy but did not identify allergies to be the cause, and Katz et al. (4) concluded that diet changes did not improve symptoms of their patients with EG.

Eosinophilic gastroenteritis is more often found in young adults than in children (5); however, the literature reveals occasional cases involving infants and children (3,4,6–9). Here, we report a case involving a 14-year-old girl who presented with acute gastric perforation and was subsequently found to have gastric ulcers based on upper endoscopy, in addition to marked eosinophilic infiltration of the stomach and duodenum. The institutional review board approved the use of the medical record for this report.


The patient was an adolescent, previously healthy girl with a 2-week history of epigastric and left upper quadrant abdominal pain. She had acute exacerbation of abdominal pain while running, with radiation of the pain to her left shoulder. The pain improved spontaneously, and she went on a canoeing trip with her family the next day. During the canoe trip, her epigastric pain worsened, and she also developed a low-grade fever and vomiting. She was subsequently seen at her local hospital's emergency department, where both an acute abdominal radiograph and computed tomography scan of her abdomen and pelvis revealed large amounts of free intraperitoneal air and fluid and thickening of the gastric wall. She was then referred to our tertiary care facility for further management.

Upon admission to our institution, her physical examination revealed a blood pressure of 104/56 mmHg, temperature of 38.2°C, pulse rate of 117/min and respiratory rate of 18/min. Head, eyes, ears, nose, throat and neck examinations were unremarkable; her chest was clear to auscultation, and heart sounds were normal. Her abdomen was mildly distended with diminished bowel sounds and with diffuse tenderness, severe diffuse peritoneal irritation and involuntary guarding. The neurological examination was grossly normal. She had no known drug allergies, and she denied any use of alcohol, tobacco or recreational drugs. The complete blood cell count at presentation revealed hemoglobin level of 13.7 g/dL, hematocrit of 0.381, platelet count of 215 K/μL, white blood count of 11.6 K/μL, with a differential of 89% neutrophils (reference range, 40%–77%), 0.7% eosinophils (reference range, 1%–7%), 1.3% monocytes (reference range, 4%–9%) and 8.2% lymphocytes (reference range, 16%–44%). Serum amylase was 75 U/L, (reference range, 30–100 U/L); total bilirubin was 0.3 (reference range, 0.4–2 mg/dL). Aspartate aminotransferase was 24 (reference range, 14–37), and alkaline phosphatase was 184 U/L (83 to 382 U/L normal). An abdominal computed tomographic scan and plain abdominal films revealed a large amount of free intraperitoneal air. She subsequently underwent an exploratory laparotomy that revealed a perforated gastric ulcer on the midbody of the stomach with no other abnormalities. The gastric perforation was repaired, and the histological examination of the initial surgical specimen revealed necrotic t with acute inflammatory exudates. The result of Helicobacter immunostain was negative.

The pediatric surgery team initiated empirical Helicobacter pylori triple therapy, followed by esomeprazole therapy, and referred the patient for further outpatient evaluation by a pediatric gastroenterologist.

The patient underwent a diagnostic esophagogastroduodenoscopy 6 weeks postsurgery which revealed 2 gastric ulcers located in the body and antrum, respectively, and a normal-appearing esophagus and duodenum. The gastric ulcers had prominent raised edges with deep erythematous central depressions and grossly appeared to be in the healing phase. There was no lymphonodular hyperplasia of the gastric antrum. Biopsies obtained from the edges of the gastric ulcers and duodenum revealed a prominent increase in eosinophilic infiltrates involving the stomach (>100 per high-power field) and, to a lesser extent, the duodenum (average, 90 per high-power field) (Figs. 1, 2). Rapid urease test (Camplyobacter-like organism test) and immunohistochemical stain were negative for Helicobacter organisms. The differential pathological diagnosis included Crohn disease, allergic gastroenteritis, idiopathic eosinophilic gastroduodenitis, drug reaction or parasitic infection.

FIG. 1
FIG. 1:
Gastric biopsy showing dense infiltration of the lamina propria with eosinophils (hematoxylin-eosin, original magnification × 20).
FIG. 2
FIG. 2:
High-power view of the gastric biopsy showing dense infiltration of a gastric pit by eosinophils (hematoxylin-eosin, original magnification × 60).

Subsequent evaluation included a normal hemoglobin level, platelet count and total white blood cell count of 6.55 K/μL; however, the differential showed 24.2% eosinophils (reference range, 1%–4%), 5.1% monocytes (reference range, 3%–9%), 40.2% lymphocytes (reference range, 33%–48%), 29.6% neutrophils (reference range, 40%–59%) and 0.9% basophils (reference range, 0%–1%). Erythrocyte sedimentation rate was 2 mm/h (reference range, 0–20 mm/h). Because the patient had not been taking any medications before her initial presentation, the possibility of a drug reaction as the etiology was excluded. The result of radioallergosorbent testing and the serological profile of the inflammatory bowel disease were negative. Stool testing was also negative for parasitic infestation.

After we reviewed the endoscopic biopsies, we started the patient on oral disodium cromoglycate, a proton pump inhibitor (PPI) and oral prednisone daily for 1 month with a taper regimen for the next month; then, the prednisone was discontinued. She remained on oral disodium cromoglycate and PPI therapy during her follow-up endoscopies. Because there is no specific protocol for the management of pediatric EG, a repeat esophagogastroduodenoscopy was performed approximately 2 weeks after the initial steroid therapy to assess histological response. Because idiopathic eosinophilic infiltration of the GI tract can also affect the colon, a colonoscopy was included to rule out involvement of the colon and terminal ileum. The colon and terminal ileum were endoscopically normal, and the biopsies were also normal.

After her repeat esophagogastroduodenoscopy revealed moderate but improved EG (up to 50 eosinophils per high-power field in the gastric biopsy), she had another 8-week tapering course of steroids. She also remained on a PPI therapy and oral disodium cromoglycate and then had a follow-up esophagogastroduodenoscopy approximately 2 weeks after completing her second steroid course. The third esophagogastroduodenoscopy, 10 months after her initial presentation, was grossly normal in appearance; duodenal biopsies were normal, and the gastric biopsies had only mild eosinophilic infiltration (up to 25 eosinophils per high-power field).

Current Status of the Patient

Her clinical outcome was satisfactory, and as mentioned previously, her subsequent esophagogastroduodenoscopies and colonoscopy revealed healed gastric ulcers and histological improvement in the degree of eosinophilic infiltration of the stomach and duodenum. Since her last esophagogastroduodenoscopy, the patient had remained clinically asymptomatic; PPI therapy was discontinued, and she remained only on oral disodium cromoglycate.


Eosinophils are normally present in varying degrees in all parts of the GI tract, and the stomach has the lowest concentration of eosinophils (10). Eosinophilic infiltration of the GI tract occurs in several diseases including the primary eosinophilic disorders such as esophagitis, gastritis, gastroenteritis and colitis and the secondary disorders that include immunoglobulin (Ig) E–mediated food allergy, gastroesophageal reflux disease, parasite infestation and inflammatory bowel disease—specifically Crohn disease.

Gastrointestinal perforations, in general, are rare in pediatric patients, and the few reported cases are either idiopathic or secondary to hollow viscus perforation. Gastrointestinal perforations secondary to EG are even rarer; as far as we know, only approximately 12 cases have been reported to date, with 7 cases reported in adults and 5 in children (6,8,11,12). Siahanidou et al. (8) reported a case of EG in a neonate which was complicated by gastric outlet obstruction, perforation of the antral wall and ileoileal intussusception. However, based on the histological findings, peripheral eosinophilia and elevated serum IgE and specific IgE to cow's milk protein, the EG, in this case, was thought to be caused by sensitivity to cow's milk protein. Furthermore, unlike our patient, the neonate in the case report also had marked infiltration of the colon, and the authors indicated that the term allergic eosinophilic gastroenterocolitis could be applied to the patient (8).

Although EG is considered to be rare in pediatrics, the eosinophilic disorders of the GI tract are being increasingly described. Reports indicate that the peak incidence of EG is in the second and third decades (13), which fits the age of presentation of our patient. The pathogenesis of EG is unknown, but postulated etiologic basis include both IgE-mediated and IgE-independent mechanisms, cytokine-mediated eosinophilic recruitment and activation (14) and T cell–mediated production of eosinophil chemokines (15).

The clinical presentation of EG is varied and is usually dependent on the affected segment of bowel. In children, EG involves the small intestine more commonly than the stomach, with the jejunum being the most common site (6,9). In contrast, in adults, the stomach was the most common site of involvement (6). The colon was the least affected segment in both children and adults (6,9). Involvement of the mucosal surface, which is most common, often leads to vomiting, abdominal pain, diarrhea, bloody or guaiac-positive stools, iron-deficiency anemia, protein-losing enteropathy and failure to thrive. Involvement of the muscularis mucosae may lead to obstructive symptoms, whereas the serosal form usually results in exudative ascites. Although no specific areas of involvement have been described for perforated lesions, it is obvious that these lesions must involve all the layers of the wall of the affected part of the GI tract. It can be speculated that the initial lesions are most likely mucosal, followed by ulcer crater formation, then infiltration of the muscularis mucosae and finally rupture through the serosal surface. Siahanidou et al. (8) postulated that the gastric perforation in their patient could be attributed to the eosinophilic infiltration of the mucosa, possibly resulting in the development of a peptic ulcer and/or the consequence of increased intraluminal pressure. The perforation in our patient was most likely secondary to an ulcer that infiltrated through the gastric wall, and the initial esophagogastroduodenoscopy revealed gastric ulcers, which seemed to be infiltrative, although this was hard to assess endoscopically.

There are no fixed criteria for the diagnosis of EG, but the diagnosis is usually based on a combination of clinical presentation, endoscopic findings, excessive eosinophilic infiltration in biopsy or resected segments, peripheral blood eosinophilia and exclusion of other causes of eosinophilic infiltration (13). The diagnosis of EG in our patient was based on the clinical presentation and biopsy findings on peripheral blood eosinophilia occurring in the postsurgical period and exclusion of other differential diagnoses. The initial surgical biopsy did not reveal significant eosinophilia because it consisted of remnants of the ulcer crater and therefore consisted mostly of necrotic t. Ureles et al. (16) proposed a descriptive classification system based on a pathological and clinical picture, with primarily diffuse versus circumscribed involvement as the main categories. Hoefer et al. (6) reported 4 cases of pediatric EG in the surgical literature who presented with a prepyloric antral web, ileal perforation, colocolic fistula and partial ileal obstruction. Interestingly, peripheral blood eosinophilia occurred in one of those cases in the postsurgical period, as in our patient. This finding is consistent with the observation that peripheral eosinophilia may be present in 80% to 100% of cases but is not pathognomonic of the condition (16). Helicobacter pylori is a more common cause of gastric and duodenal ulcers, but the results of Helicobacter immunostain of the resected surgical t (and subsequent biopsies) was negative, suggesting that the ulcers in our patient were not caused by H. pylori infection. Furthermore, H. pylori infection is not a known cause of EG, and the eosinophilic infiltration would have been expected to resolve after appropriate H. pylori therapy. The degree of eosinophilic infiltration of the stomach and duodenum was suggestive of a short-term process that most likely was present at the patient's initial presentation.

Unlike eosinophilic esophagitis, the management of EG is not well established. Dietary elimination and or elemental diets have been used in cases in which specific foods have been implicated in the etiology of the disease. Other treatment modalities include mast cell stabilization agents such as oral disodium cromoglycate, ketotifen (H1 antihistamine) and montelukast (leukotriene antagonist) and corticosteroids (13). Our patient was treated with a combination of oral disodium cromoglycate, PPI and corticosteroids, with good clinical and histological response.

Our case is unusual in terms of the acute nature of the presentation and in the severity of the initial presentation. Because the primary eosinophilic disorders of the GI tract are being increasingly described in pediatric gastroenterology practice, it is important to recognize the diverse nature of this pathology. This case illustrates the diverse presentation of EG in children, including gastric perforation (without any chronic antecedent symptoms) as a possible complication. Furthermore, the patient showed good clinical and histological response to mast cell stabilizer, PPI and corticosteroid therapy.


The authors thank Suzanne Speaker for assistance with manuscript preparation.


1. Leinbach GE, Rubin CE. Eosinophilic gastroenteritis: a simple reaction to food antigens? Gastroenterology 1970; 59:874–889.
2. Hummer-Ehret BH, Rohrschneider WK, Oleszczuk-Raschke K, Darge K, Nutzenadel W, Troger J. Eosinophilic gastroenteritis mimicking idiopathic hypertrophic pyloric stenosis. Pediatr Radiol 1998; 28:711–713.
3. Caldwell JH, Sharma HM, Hurtubise PE. Eosinophilic gastroenteritis in extreme allergy. Gastroenterology 1979; 77:560–564.
4. Katz AJ, Twarog FJ, Zeiger RS, Falchuk ZM. Milk-sensitive and eosinophilic gastroenteropathy: similar clinical features with contrasting mechanisms and clinical course. J Allergy Clin Immunol 1984; 74:72–78.
5. Trounce JQ, Tanner MS. Eosinophilic gastroenteritis. Arch Dis Child 1985; 60:1186–1188.
6. Hoefer RA, Ziegler MM, Koop CE, Schnaufer L. Surgical manifestations of eosinophilic gastroenteritis in the pediatric patient. J Pediatr Surg 1977; 12:955–962.
7. Maluenda 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; 3:349–356.
8. Siahanidou T, Mandyla H, Dimitriadis D, Van-Vliet C, Anagnostakis D. Eosinophilic gastroenteritis complicated with perforation and intussusception in a neonate. J Pediatr Gastroenterol Nutr 2001; 32:335–337.
9. Whitington PF, Whitington GL. Eosinophilic gastroenteropathy in childhood. J Pediatr Gastroenterol Nutr 1988; 7:379–385.
10. Lowichik A, Weinberg AG. A quantitative evaluation of mucosal eosinophils in the pediatric gastrointestinal tract. Mod Pathol 1996; 9:110–114.
11. Agertoft A, Husby S, Host A. Intestinal perforation in a two-year-old child with eosinophilic gastroenteritis. Acta Paediatr Scand 1991; 80:389–391.
12. Deslandres C, Russo P, Gould P, Hardy P. Perforated duodenal ulcer in a pediatric patient with eosinophilic gastroenteritis. Can J Gastroenterol 1997; 11:208–212.
13. Kirschner B. Undetermined colitis and other inflammatory diseases. In: Walker W, Goulet O, Kleinman RE, Sherman PM, Shneider BL, Sanderson IR, eds. Pediatric Gastrointestinal Disease: Pathophysiology, Diagnosis, Management, 3rd ed. Ontario, CA: BC Decker Inc, 2004:850-2.
14. Desreumaux P, Bloget F, Seguy D, et al. Interleukin 3, granulocyte-macrophage colony-stimulating factor, and interleukin 5 in eosinophilic gastroenteritis. Gastroenterology 1996; 110:768–774.
15. Teitelbaum JE, Fox VL, Twarog FJ, et al. Eosinophilic esophagitis in children: immunopathological analysis and response to fluticasone propionate. Gastroenterology 2002; 122:1216–1225.
16. Ureles A, Alschibaja T, Lodico D, Stabins SJ. Idiopathic eosinophilic infiltration of the gastrointestinal tract, diffuse and circumscribed: a proposed classification and review of the literature, with two additional cases. Am J Med 1961; 30:899–909.
© 2006 Lippincott Williams & Wilkins, Inc.