This article is accompanied by an editorial. Please see: Furuta GT. Eosinophils in the esophagus: acid is not the only cause. J Pediatr Gastroenterol Nutr 1998;26:468-471.
In 1982, Winter et al. correlated the histologic appearance of esophageal eosinophils with abnormal acid clearance (1). Since then, the presence of esophageal eosinophils without evidence of other gastrointestinal disorder has been the hallmark of gastroesophageal reflux and esophagitis in children. Although Winter et al. suggested that the number of esophageal eosinophils increases in proportion to the amount of acid emitted in the distal esophagus, in most cases of reflux esophagitis, only a small number of esophageal eosinophils are present (counted per high-power microscopic field [HPF]).
In the past, patients with chronic symptoms of gastroesophageal reflux(GER) and esophageal eosinophilia that persisted despite medical therapy were often referred to a surgeon for an antireflux procedure(2,3). Recently, several reports in the adult literature have described patients with idiopathic eosinophilic esophagitis with persistent dysphagia, esophageal spasm, or esophageal stricture, with the eosinophilic esophagitis responding to steroids(4-6). Kelly et al. have reported a series of 10 children with symptoms of GER and eosinophilic esophagitis who improved with the introduction of an amino acid-based formula (7).
We report a series of 20 children with symptoms of GER and primary eosinophilic esophagitis that were unresponsive to aggressive antireflux medical therapy but that responded both clinically and histologically to oral corticosteroids.
MATERIALS AND METHODS
Between January 1, 1993, and July 1, 1995, 1809 patients were evaluated for GER disease. Each patient displayed chronic gastrointestinal symptoms (>2 months' duration) including abdominal-epigastric pain or regurgitation-vomiting, and at least one of the following: nausea, globus, water brash, chest pain, dysphagia, nighttime coughing, choking, poor appetite, weight loss, or irritability. Upper gastrointestinal series were performed in all patients. Whenever symptoms persisted or recurred despite medical management (positioning, feeding alteration, antacids, or H2 blockers) or when the symptoms were complicated by gastrointestinal bleeding, respiratory disease, or weight loss, an esophagogastroduodenoscopy (EGD) was performed by a board-certified pediatric gastroenterologist using an Olympus video endoscope (Olympus, Columbia, MD, U.S.A.) (N30, XP20, or GIF 100-depending on the patient's age). Patients with known gastrointestinal disorders (Crohn's disease, ulcerative colitis, celiac disease), anatomic abnormalities (malrotation, hiatal hernia, duplication) or systemic disease(cancer, chronic renal disease, scleroderma) were excluded.
Five hundred eighty-three patients underwent EGD for GER. Of this group, reflux esophagitis was diagnosed and treated in 418. Initial treatment consisted of ranitidine with metoclopramide or cisapride. Those patients who improved during the therapy were considered to be the reference group of children with GER. If the patients' symptoms worsened or did not show at least a 50% improvement with medication, the dose of ranitidine was increased to a maximum of 3 mg/kg twice daily, and patients were given cisapride at a maximum dose of 0.3 mg/kg four times daily. After 3 months, another EGD was performed if symptoms remained despite the therapy. Patients with continued histologic evidence of esophagitis continued receiving cisapride but with the addition of omeprazole (1 mg/kg per day; minimum, 10 mg/day; maximum, 20 mg twice daily).
Patients who remained symptomatic despite more than 3 months of omeprazole and cisapride therapy again underwent EGD (11.2 ± 3.8 months after the initial EGD). Those in whom continued severe esophageal eosinophilia was demonstrated, consisting of more than 15 eosinophils/HPF without evidence of antral or duodenal eosinophilia were defined as the study population. In patients with antral or duodenal eosinophilia, in addition to esophageal eosinophilia, eosinophilic gastroenteritis was diagnosed, and the patients were excluded from the study. Figure 1 outlines selection of the study population.
At least two random grasp biopsies were taken of the duodenum (third portion), stomach (antrum), and esophagus (3-5 cm above Z-line) in every patient. Additional biopsies were taken from any other site that was visually abnormal. The average number of biopsy specimens taken from each site was 2.3± 1.1. Each specimen was evaluated for inflammatory infiltrate(neutrophils, eosinophils) and ulceration. In addition, duodenal specimens were evaluated for villous and crypt abnormalities, antral specimens for the presence of Helicobacter pylori, and esophageal specimens for reflux esophagitis (8,9). In every esophageal specimen, the number of eosinophils per HPF (magnification, ×40) was determined by averaging the number of eosinophils in sequential HPFs over the entire area(one level) of the specimen.
All study patients underwent 24-hour pH probe testing (to determine whether acid reflux was related to esophageal eosinophilia) and laboratory evaluation, including complete blood count with differential, total serum eosinophil count, quantitative immunoglobulin (Ig) E level, chemistry panel, and sedimentation rate. Patients with severe reflux revealed by pH probe (reflux index >40%) underwent surgical evaluation for fundoplication. The remaining patients began a 4-week course of oral methylprednisolone. Before therapy, each patient and family member was interviewed and completed a questionnaire regarding the type, chronicity, and frequency of symptoms (including emesis, regurgitation, heartburn, globus, dysphagia, irritability, weight loss, anorexia, water brash). Factors that precipitated symptoms-foods, position, and exercise-were also recorded. Each week, a telephone interview was conducted by the physician to record any change in clinical symptoms. All acid-suppressing medication and prokinetic agents were continued. After 4 weeks of therapy, a second upper endoscopic study, complete blood count, total serum eosinophil count, and quantitative IgE were performed.
Comparison of esophageal eosinophilia, total eosinophil count, and quantitative IgE was performed by the Wilcoxon matched-pairs signed ranks test(SPSS for Windows, Chicago, IL, U.S.A.).
An EGD was performed in 583 patients: Findings were normal in 165, whereas 418 had histologic evidence of esophagitis. Of these, 214 had at least 1 esophageal eosinophil/HPF without evidence of gastric or duodenal disease(Figure 1). After treatment with ranitidine-omeprazole and metoclopramide-cisapride, clinical symptoms improved in 184 patients (control group). The clinical characteristics of the control group consisted of 94% with regurgitation-vomiting, 82% with abdominal-epigastric pain, 69% with nausea, 53% with chest pain-heartburn, 22% with weight loss, 17% with water brash-globus, 16% with choking, 13% with dysphagia, and 8% with bronchospasm, coughing, and eczema.
Findings in a second EGD examination showed 30 patients with continued esophageal eosinophilia (>15 eosinophils/HPF) and persistent symptoms. In 8 patients with antral or duodenal eosinophilia (in addition to their continued esophageal eosinophilia) eosinophilic gastroenteritis was diagnosed and the patients excluded from the study. The remaining 22 patients underwent a 24-hour pH probe (Synectios pH probe, Irving, TX, U.S.A.), with all acid-suppressing agents and prokinetic medications withdrawn. Two patients had severe GER according to pH probe criteria (reflux index >40%; total time and longest episodes >90 minutes) and underwent an antireflux procedure. Twenty patients demonstrated mild or no reflux in examination by pH probe but continued to have severe symptoms and eosinophilic esophagitis, despite medical therapy with omeprazole and cisapride (Table 1).Table 2 depicts the 24-hour pH probe characteristics of the study population. The average duration of symptoms in these patients was 3.2 ± 1.8 years. Results of laboratory testing revealed 1 patient with an elevated sedimentation rate and 1 with hypoalbuminemia.
The average number of esophageal eosinophils/HPF was significantly greater in those children with eosinophilic gastroenteritis and primary eosinophilic esophagitis (34.7 ± 17.3 eosinophils/HPF and 34.2 ± 9.6 eosinophils/HPF, respectively), compared with the count in children with GER that responded to antireflux therapy (2.26 ± 1.16 eosinophils/HPF;p < 0.001); however, no statistical difference was found in the average number of esophageal eosinophils/HPF between children with primary eosinophilic esophagitis and those with eosinophilic gastroenteritis.
Results of Corticosteroid Therapy
Twenty patients entered the study and began a 4-week course of 1.5 mg/kg methylprednisolone twice daily. Thirteen of 20 patients became completely asymptomatic and 6 others had marked improvement in their clinical symptoms after 4 weeks of corticosteroid therapy. The average time for initial appearance of clinical improvement was 8 ± 4 days. One patient continued to experience abdominal pain similar to the pain described before the initiation of steroid therapy. All 20 patients demonstrated significant clinical (asymptomatic) improvement, which was supported by histologic evidence of the average number of esophageal eosinophils/HPF. In addition, a significant difference was seen in the serum eosinophil count and quantitative IgE level (Table 3). Figure 2 shows the esophageal histologic improvement after corticosteroids.
LONG-TERM FOLLOW UP
After a second EGD, each patient's corticosteroids and reflux medications were tapered and withdrawn in 6 weeks. At the 12-month follow up, 10 patients(50%) remained asymptomatic. In 9, symptoms redeveloped and dietary restrictions were imposed, as previously described (7). Of these patients, 7 (35%) responded to dietary withdrawal, whereas 2 (10%) remained symptomatic and required a second corticosteroid challenge, 6± 2.5 months after initial steroid therapy. One patient (5%) remained symptomatic despite corticosteroids and dietary restriction.
These results demonstrate that a significant number of children with characteristic symptoms of gastroesophageal reflux and predominant eosinophilic esophagitis do not have GER but instead have a unique disorder, which does not respond to antireflux therapy but does respond both clinically and histologically to oral corticosteroids. The features of this disorder, observed in the patients reported here, are significant eosinophilic infiltration only of the esophagus; minimal-to-no acid reflux detectable by 24-hour pH monitoring; and resolution of the esophageal eosinophilia, evident in histologic study, within 1 month of initiation of steroid therapy. Although not statistically significant, the clinical features of dysphagia, eczema, and bronchospasm occurred more often in the study group, compared with occurrences of those symptoms in the control group with reflux esophagitis. After administration of corticosteroids, rapid improvement in clinical symptoms was observed in 19 of the 20 patients (noted within as few as 4 days); histologic evidence of improvement was documented within 4 weeks of therapy.
The precise role of the esophageal eosinophil has not been defined. Esophageal eosinophilia was linked to pediatric reflux esophagitis in 1982 by Winter et al. (1) They correlated the observation of esophageal eosinophils with abnormal esophageal acid clearance. However, none of the 46 patients studied had more than 3 eosinophils per HPF. Since that time, in several reports adult patients with severe eosinophilic esophagitis have been identified, suggesting a cause other than acid reflux. In 1985, Lee reported on a series of 11 patients with more than 10 esophageal eosinophils per HPF who had dysphagia, heartburn, vomiting, and esophageal strictures (in 3 of 11) (10). Although he suggested that reflux occurred in the majority, reflux was not indicated by results of 24-hour pH probe. One patient was given steroids and showed clinical improvement.
In 1993, Attwood et al. described 12 patients with dysphagia who had more than 20 esophageal eosinophils/HPF (mean, 56/HPF) (11). Normal findings in pH monitoring were noted in 92%, whereas 56% had evidence of an allergic disorder. He compared results in this group with those in patients with medically responsive GER (documented by 24-hour pH probe) of whom only 45% had eosinophils in the esophagus and to a much lesser degree(mean, 3.3 eosinophils/HPF). Vitellas et al. reported 13 male patients with primary esophageal eosinophilia, dysphagia (n = 12), allergic manifestations (n = 10), peripheral eosinophilia (n = 12) and proximal esophageal strictures (n = 10)(12). None had reflux observed in barium contrast study and 85% improved with steroid therapy. In one patient, results of a second barium contrast esophagram showing resolution of the stricture.
Although all of our study patients responded rapidly to corticosteroids and 50% remained asymptomatic 12 months after therapy, allergic enteritis should always be considered in these patients. A food elimination and challenge test should be conducted when parents or physicians choose not to use corticosteroids or when symptoms recur. Kelly et al. reported on a series of children with esophageal eosinophils who did not respond to antireflux therapy but did respond to an amino acid-based formula (7). Only 1 patient underwent a 24-hour pH probe, in which the results showed no evidence of reflux. The patients were given solely an an amino-acid based formula for a median of 17 weeks. Symptomatic improvement was seen within an average of 3 weeks; subsequently, regular foods were slowly reintroduced. Kelly et al. suggested an immunologic basis secondary to delayed hypersensitivity or to a cell-mediated hypersensitivity response.
Our results demonstrate that eosinophilic esophagitis may take months to years to evolve. One fourth of our patients did not demonstrate severe esophageal eosinophilia on initial endoscopic examination. In each of these patients, pronounced and persistent esophageal eosinophilia was found in a second EGD after failure to obtain a clinical response to antireflux medication. Similarly, whereas 70% of patients demonstrated peripheral eosinophilia, 35% did not until 5 months after initial evaluation. In 26%, patients with isolated esophageal eosinophilia had evidence of antral or duodenal eosinophilia in a second endoscopic study (eosinophilic gastroenteritis).
In several reports, it has been suggested that eosinophilic gastroenteritis can involve the esophagus (13-15). Eosinophilic gastroenteritis is a poorly characterized, chronic disorder in which eosinophils infiltrate various layers of the alimentary tract, causing myriad gastrointestinal symptoms, including abdominal pain, vomiting, diarrhea, gastrointestinal bleeding, protein losing enteropathy, weight loss, and failure to thrive (16). In the past, the diagnosis of eosinophilic enteritis was not made unless eosinophils were observed in another portion of the gastrointestinal tract. Eosinophilic tissue infiltration in eosinophilic gastroenteritis can occur in all layers and in every part of the gastrointestinal tract (17). In children, findings of antral or duodenal eosinophilia, coupled with peripheral eosinophilia, suggest the diagnosis. Other gastrointestinal sites, including the esophagus and colon, have been reported to be affected in eosinophilic gastroenteritis but always in association with gastric or small bowel disease(14). The majority of patients manifest peripheral eosinophilia; an elevated IgE level; and other allergic symptoms, including asthma, rhinitis, and eczema.
Although the findings in this study do not answer the question of cause, they indicate that children with presumed severe reflux esophagitis that is unresponsive to aggressive antireflux medical therapy should not be immediately referred for antireflux surgery but instead should be evaluated for other immunologic and allergic causes of the esophageal eosinophilia. This study would have benefitted from inclusion of a placebo-controlled trial; however, the ethical considerations of using a placebo in children with severe clinical and histologic symptoms far outweighed the desirability of such a trial. Future research should be performed to determine the correlation of the number of esophageal eosinophils to specific disease (neonatal gastroesophageal reflux, recurrent reflux in infants and older children, allergic esophagitis and immunologic-based disease). In addition, research should be performed to determine the role of the enteric eosinophil and the cause of primary eosinophilic esophagitis.
We believe that children in whom gastroesophageal reflux and esophagitis are diagnosed, who do not respond to aggressive antireflux therapy, should undergo another round of testing for primary eosinophilic esophagitis and for the development of eosinophilic gastroenteritis. If significant esophageal eosinophilia is noted on histologic study of a biopsy specimen, or if peripheral eosinophilia develops, primary eosinophilic esophagitis should be considered. Steroid therapy may offer advantages as the initial therapy; our results demonstrated rapid clinical and histologic improvement while the child consumed a diet of conventional foods.
1. Winter HS, Madara JL, Stafford RJ. Intraepithelial eosinophils: A new diagnostic criterion for reflux esophagitis.Gastroenterology
2. Schatzlein MH, Ballantine TWN, Thirunavukkarasu S, et al. Gastroesophageal reflux in infants and children. Arch Surg
3. Hill LD, Aye RW, Ramel S. Antireflux surgery: A surgeon's look. Gastroenterol Clin North Am
4. Attwood SEA, Smyrk TC, DeMeester TR, et al. Esophageal eosinophilia with dysphagia: A clinico pathologic syndrome. Dig Dis Sci
5. Vitellas KM, Bennett WF, Bova JG, et al. Idiopathic eosinophilic esophagitis. Radiology
6. Levine MS, Saul SH. Idiopathic eosinophilic esophagitis: How common is it? Radiology
7. Kelly KJ, Lazenby AJ, Rowe PC, et al. Eosinophilic esophagitis attributed to gastroesophageal reflux: Improvement with an amino acid-based formula. Gastroenterology
8. Behar J, Sheahan DC. Histologic abnormalities in reflux esophagitis. Arch Pathol Lab Med
9. Cucchiara S, D'Armiento F, Alfieri E, et al. Intraepithelial cells with irregular nuclear contours as a marker of esophagitis in children with gastroesophageal reflux disease. Dig Dis Sci
10. Lee RG. Marked eosinophilia in esophageal mucosal biopsies. Am J Surg Pathol
11. Attwood SEA, Smyrk TC, Demeester TR, Jones JB. Esophageal eosinophilia with dysphagia: A distinct clinicopathologic syndrome.Dig Dis Sci
12. Vitellas KM, Bennett WF, Bova JG, et al. Idiopathic eosinophilic esophagitis. Radiology
13. Dobbins JW, Sheahan DG, Behar J. Eosinophilic gastroenteritis with esophageal involvement. Gastroenterology
14. Matzinger MA, Daneman A. Esophageal involvement in eosinophilic gastroenteritis. Pediatr Radiol
15. Goldman H, Proujansky R. Allergic proctitis and gastroenteritis in children: Clinical and mucosal biopsy features in 53 cases.Am J Surg Pathol
16. Whitington PF, Whitington GL. Eosinophilic gastroenteropathy in childhood. J Pediatr Gastroenterol Nutr
17. Klein NC, Hargrove RL, Sleisinger MH, Jeffries GH. Eosinophilic gastroenteritis. Medicine
In the Next Issue
Celiac Disease and Turner Syndrome
M. Bonamico, G. Bottaro, A. M. Pasquino, M. Caruso-Nicoletti, P. Mariani, G. Gemme, E. Paradiso, M. C. Ragusa, and M. Spina
Serum Leptin in Children and Young Adults with Inflammatory Bowel Disease
Alison G. Hoppin, Lee M. Kaplan, David Zurakowski, Alan M. Leichtner, and Athos Bousvaros
The Effect of Intestinal Permeability on Pancreatic Enzyme-Induced Enteropathy in the Rat
John D. Lloyd-Still, Michael R. Uhing, Valerie Arango, Aldo Fusaro, and Robert E. Kimura
Efficacy of a Glutamine-Based Oral Rehydration Solution on the Electrolyte and Water Absorption in a Rabbit Model of Secretory Diarrhea Induced by Cholera Toxin
Alessandra C. Silva, Messias S. Santos-Neto, Alberto M. Soares, Manassés C. Fonteles, Richard L. Guerrant, and Aldo A. M. Lima
Epithelial Growth of the Small Intestine in Human Infants
Fiona M. Thompson, Anthony G. Catto-Smith, David Moore, Geoff Davidson, and Adrian G. Cummins
Glutamine-Based Oral Rehydration Solutions: The Magic Bullet Revisited?