GASTROENTEROLOGY 1995;109:1503-1512 Eosinophilic Esophagitis Attributed to Gastroesophageal Reflux: Improvement With an Amino Acid-Based Formula KEVIN J. KELLY, *'~ AUDREY J. LAZENBY, g PETER C. ROWE,* JOHN H. YARDLEY, II JAY A. PERMAN,*'* and HUGH A. SAMPSON *'~ Divisions of *Pediatric Gastroenterology/Nutrition and ~PediatricAllergy/Immunology and Departments of *Pediatrics and EiPathology, The Johns Hopkins UniversitySchool of Medicine, Baltimore, Maryland; and gDepartment of Pathology, Universityof Alabama at Birmingham, Birmingham, Alabama Background & Aims: Treatment for gastroesophageal reflux may be ineffective in patients with an eosino- philic infiltration of the esophagus. The aim of this study was to investigate whether unremitting symp- toms of gastroesophageal reflux and biopsy abnormali- ties of the esophagus may be associated with the in- gestion of certain foods. Methods: Ten children pre- viously diagnosed with gastroesophageal reflux by standard testing with long-standing symptoms (me- dian, 34.3 months; range, 6-78 months) despite stan- dard antireflux therapies, including Nissen fundoplica- tion in 6 patients, were fed the elemental formulas Neocate or Neocate-l-Plus (Scientific Hospital Sup- plies Inc., Gaithersburg, MD) for a minimum of 6 weeks. Each child had repeat endoscopy followed by open food challenges. Results: While receiving the formulas, pa- tients had either resolution (n = 8) or improvement (n -- 2) of symptoms. On follow-up esophageal biopsy, the maximal intraepithelial eosinophil counts de- creased significantly before (median, 41; range, 15- 100) to after (median, 0.5; range, 0-22) the formula trial (P = 0.005). Other reactive epithelial changes of the esophageal mucosa also improved significantly. All patients redeveloped their previous symptoms on open food challenges. Conclusions: Chronic gastrointestinal symptoms and histological changes of the esophagus unresponsive to standard treatments for gastroesopha- geal reflux were improved by the use of elemental for- mulas. Symptoms recurred when specific dietary pro- teins were reintroduced during open food challenges. The mechanism of these observations is unknown. T he presence of eosinophils in esophageal mucosal biopsy specimens is considered a highly specific marker for gastric acid reflux. In 1982, Winter et al. first correlated delayed clearance of acid from the esophagus measured by intraesophageal pH probe monitoring with an eosinophilic infiltration of the esophageal mucosa. I Esophageal eosinophils were also correlated with basal- zone hyperplasia and elongation of the vascular papillae, two findings considered representative of reactive changes of the esophageal epithelium to acid reflux. 2-4 Winter et al. concluded that the eosinophilic infiltration was also an early histological sign of acid reflux and that the more prominent the infiltration, the greater the duration or severity of the reflux. Since that report, an eosinophilic infiltration of the esophageal mucosa with or without basal-zone hyperplasia and elongated vascular papillae has been consistently interpreted as diagnostic of gastro- esophageal reflux disease (GERD)) -7 We recently evaluated 10 pediatric patients with in- tractable symptoms attributed to GERD who also had long-term histological changes of the esophageal mucosa, including basal-zone hyperplasia, elongation of vascular papillae, and an eosinophilic infiltration. All 10 patients had the diagnosis of GERD made previously by one or more abnormal results of tests for reflux. Each patient had been compliant with several courses of standard antireflux medical therapies, although no treatment had resulted in the resolution of either the long-term symptoms or the esophageal mucosal abnormalities. Ultimately, 6 children had a Nissen fundoplication. To explain these treatment failures, we proposed the following alternative hypothesis: the intractable symp- toms and the esophageal eosinophilic infiltration were present not as a result of persistent acid reflux but rather as a result of a response to the ingestion of intact dietary proteins. We speculated that this response may be a hypersensitivity reaction. We further speculated that these patients may have regularly eaten these dietary proteins without an awareness of the relationship be- tween the ingestion of the foods and the development of symptoms. In previous studies, elimination of individual or multi- Abbreviations used in this paper: GERD, gastroesophageal reflux disease. © 1995 by the AmericanGastroenterological Association 0016-5085/95/$3.00