Elemental Diet Improves Steroid-Dependent Eosinophilic Gastroenteritis and Reverses Growth Failure

Abstract

Eosinophilic gastroenteropathy (EGE) is an uncommon disorder characterized by eosinophilic infiltration of the intestine, a variety of gastrointestinal symptoms, and often accompanied by peripheral eosinophilia (1-3). The course of the disease is highly variable, with some patients responding to elimination diets (3,4), whereas others require short courses of corticosteroids (2,3,5). We report the unusual case of an 18-year-old patient with multiple food allergies and severe, corticosteroid-dependent EGE. Prolonged use of an elemental diet allowed the complete withdrawal of steroids and reversed his growth failure. CASE REPORT The patient, a white boy, was first seen with failure to thrive, vomiting, and diarrhea at 6 months of age. At 11 months, hypoalbuminemia (albumin, 25.2 g/L) secondary to a protein-losing enteropathy was confirmed by 51Cr-labeled albumin loss in stools (12.1%; normal, 01.-0.7%). He was also diagnosed as having asthma and multiple food allergies [skin test and radioallergosorbent test (RAST) positive]. Peripheral blood eosinophilia was present (absolute count, 2,040). An initial jejunal biopsy revealed severe villous atrophy with mixed acute (neutrophils, eosinophils) and chronic (lymphocytes, plasma cells) inflammatory cell infiltrates of the lamina propria. Diagnoses of celiac disease and cow's milk protein-induced enteropathy were considered, and a gluten- and diet was instituted. Despite strict adherence to the diet, his failure to thrive persisted along with intermittent irritability, diarrhea, and vomiting. Multiple jejunal biopsies at various times either demonstrated mild villous atrophy with variable degrees of mixed inflammatory cell infiltrates or were completely normal. At 29 months of age, symptoms rapidly improved with an initial empiric trial of oral corticosteroids, tapered over a 1-month period. Failure to thrive persisted, however, primarily due to insufficient caloric intake. At 3 years of age, the patient had a recurrence of gastrointestinal symptoms, protein-losing enteropathy, and asthma. At this time, his absolute eosinophil count was 1,550, and the repeat jejunal biopsy was normal. However, a gastric antral biopsy showed an intense inflammatory infiltrate of the mucosa with a marked predominance of eosinophils, consistent with a diagnosis of eosinophilic gastroenteropathy (1,3). Corticosteroid therapy was rapidly effective in controlling his symptoms. However, multiple attempts to withdraw the medication over the ensuing years were unsuccessful, with exacerbation of abdominal pain, vomiting, and diarrhea. Treatment with oral Na chromoglycate did not have any beneficial effect. Corticosteroids were withdrawn completely on only one occasion, during a course of bowel rest and total parenteral nutrition. Otherwise, daily or alternate day prednisone (10-20 mg) was required to control symptoms provoked by food. Growth failure became a major concern, with an average growth velocity of only 2.8 cm/year (>2 SD below mean) between 7 and 17 years of age (Fig. 1). At 17 years, his weight was 40 kg, and height 145 cm, both significantly below the third percentile. Caloric intake was adequate for his needs [≥100% recommended daily intake (RDI)], and he was not underweight (111% of ideal weight for height). Bone age and puberty were markedly delayed (11 years and Tanner II, respectively). Monthly testosterone injections were given for a 10-month period, with no effect on growth, although his Tanner stage progressed. Before elemental diet (ED) therapy, laboratory investigations revealed peripheral eosinophilia, elevated serum immunoglobulin E (IgE), and hypoalbuminemia (Table 1). Skin-prick tests were still positive to multiple foods, including milk, eggs, chicken, lamb, turkey, and even rice. He was then treated with a strict ED using Vivonex TEN (Sandoz Nutrition) at a concentration of 1 kcal/cc. The elemental formula was administered progressively by nocturnal gavage (2,000 kcal/day), according to our protocol for Crohn's disease complicated by growth failure (6). He was permitted daytime clear fluids and flavored Vivonex drinks. His only other food intake was potatoes prepared with medium-chain triglyceride (MCT) oil, as well as iron and vitamin K supplements. During 18 consecutive months of treatment, the patient remained completely asymptomatic, and his growth velocity increased dramatically to 8.4 cm/year, whereas the Tanner stage remained unchanged. Furthermore, all laboratory parameters normalized while on ED therapy (Table 1). Repeat endoscopic biopsies after 3 months of strict ED were improved, with minimal eosinophilic infiltrates in the distal esophagus and antrum. The duodenal biopsy was much improved, with normal villus crypt architecture and only moderate infiltration of eosinophils and mast cells. He did not consent to repeat the endoscopy for re-evaluation of histology subsequently. Corticosteroids were tapered and completely discontinued 8 weeks after initiation of the ED therapy. The exclusion diet and nocturnal nasogastric feedings were very well tolerated by the patient. His asthma was clinically improved as well. Vitamins B12, folate, A, D, E, and K-dependent coagulation factors were all normal, as were calcium, magnesium, phosphate, iron, zinc, and copper. However, the treatment was complicated by deficiencies of selenium (0.3 μmol/L; normal serum values, 1.0-2.0) and essential fatty acids (ratio of 20:3ω9/20:4ω6 of 0.05, with control plasma ratio 0.016 ± 0.005), detected by GLC (7) after 10 months of ED therapy. Both abnormalities were easily corrected with oral supplementation (Se, 240 μg/day; safflower oil 15 ml/day). To test his cellular immune responsiveness to food antigens, in vitro peripheral blood mononuclear cell (PBMC) cultures were carried out in the presence of protein food antigens, before and after 16 months of ED therapy, as previously described (8). PBMC cultures were maintained for 6 days, and 3H-thymidine incorporation during the last 4 h of culture was measured in wells with antigen and expressed as a stimulation index compared with control wells with medium alone. The initially observed increased proliferation index in the presence of whey and β-lactoglobulin decreased dramatically after ED to values obtained in patients with nonallergic enteropathies (Table 2). These improvements were associated with the patient's being asymptomatic. Normal PBMC proliferation indices were observed for this patient to ovalbumin and a mitogen (PHA), irrespective of diet. Oral challenges with various food allergens had consistently been positive in the past, correlating well with skin and lymphocyte tests. Therefore, we elected to continue his ED therapy without introducing foods to achieve maximal catch-up growth. After 22 months of ED therapy, the patient decided to discontinue the ED gavage feeds and began eating nondiscriminantly. This was followed by a rapid recurrence of postprandial epigastric pain and vomiting, as well as a recurrence of asthma. Ketotifen (2 mg p.o. t.i.d.) was used without success and was abandoned because of somnolence and irritability. Over the last 3 years, his symptoms are well controlled only while receiving alternate-day prednisone (≥17.5 mg every 2 days). His weight was maintained, but growth velocity decelerated to 2 cm/year. Further ED therapy was refused. DISCUSSION Eosinophilic gastroenteropathy is a chronic, illdefined disorder of the gastrointestinal tract with nonspecific digestive symptoms, often accompanied by peripheral blood eosinophilia. The hallmark of the disease is predominant infiltration of the intestine by eosinophils (9), most often in the mucosa, but in some cases, in the muscle layer, or rarely in the serosa, making the diagnosis more difficult (10,11). The lesions may be focal, as exhibited in this case, requiring biopsies from multiple sites before a specific diagnosis can be made (12). This report describes the first successful use of prolonged ED to resolve all symptoms and signs of severe EGE, allowing cessation of steroids and reversal of growth failure. The rapid increase in growth velocity during ED treatment in our patient most likely represents the steroid-sparing effect of this therapy, as well as improving the protein-losing enteropathy and nutritional status, as we previously described in pediatric Crohn's disease (6). Caloric and protein intake were within normal limits before ED therapy, and the patient was stunted but not underweight for his height. During ED therapy, his serum protein levels normalized rapidly, consistent with reversal of the protein-losing enteropathy, as has been noted in Crohn's disease treated with ED (13). Steroid-responsive protein-losing enteropathy associated with gastrointestinal allergy was reported by Greenberger et al. (14). Although a radiolabeled albumin study was not repeated, our patient's course is consistent with a similar effect of ED, although improved nitrogen balance on ED therapy may also have contributed. Previous reports have noted variable results with elemental diet therapy for this condition. Tytgat et al. (15) described a 66-year-old man who died of an “eosinophilic enteritis” and bacterial overgrowth, who had received total parenteral nutrition and an unspecified elemental diet therapy without benefit. Nelson et al. (16) reported (in abstract form) the case of a 9-month-old girl with protein-losing enteropathy, eosinophilia, and biopsy proven EGE. She was treated with Vivonex with improvement in symptoms and biopsy findings. On the other hand, Caldwell et al. (17) described a case of atopy associated with EGE in a 13-year-old boy with exacerbation of allergic pruritis and wheezing on an elemental diet. Leinback and Rubin (4) reported a patient who initially improved with elimination of all proteins other than soy, but who relapsed after several months. The severe villous atrophy without eosinophilic infiltrates found on the initial jejunal biopsies in our patient are atypical of EGE, although the infiltrates are often focal (1-3, 12). The course of the disease, negative anti-gliadin antibodies, and the lack of response to a gluten-free diet rule out a diagnosis of celiac disease. Villous atrophy has been described with cow's milk protein allergy, as well as in EGE (1,4,18,19). The coexistence of multiple food allergies and eosinophilic gastroenteropathy was confirmed by food challenges, positive skin-prick tests, and RAST antibody levels in our patient. There is considerable controversy in the literature regarding this association (2-4,20,21). Two distinct EGE groups have been reported, one of which includes patients with atopy, asthma, high levels of serum IgE, and skin test and RAST positivity (17). Patients in this group have been shown to have relapses of symptoms, as well as increases in peripheral eosinophilia and serum IgE, in response to specific food challenges. The second group does not have a history of atopy and usually has normal IgE levels and negative skin and RAST tests. Both groups frequently require corticosteroid treatment, and the results of food allergen “elimination diets” have been variable. The dramatic response to ED therapy, with the exclusion of all intact protein from the diet and relapse when normal foods were reintroduced, clearly establishes a role of dietary protein antigens in the pathogenesis of EGE in our patient. In our recent series of 20 consecutive pediatric EGE patients, only one third of cases had features of atopy, and the majority did not respond to hypoallergenic diets (bovine and soy protein free), requiring corticosteroids to control symptoms (12). However, complete exclusion of food allergens with ED was not attempted. In addition to the evidence supporting the role of IgE-mediated allergy in this case, abnormal in vitro responsiveness of PBMCs to cow's milk protein antigens was documented, implicating cell-mediated immune mechanisms (8,12). However, other dietary antigens were likely to be important, as the patient did not respond to a strict cow's milk protein-free diet. The response to food antigens by his PBMCs in vitro suggests that his circulating immune cells are sensitized to these proteins. In atopic patients, it has been possible to generate T-helper cell clones from blast cells derived from PBMCs stimulated with specific antigens (22). These Th1 CD4+ cell clones secrete cytokines implicated in IgE and eosinophilic responses, including interleukins 4 and 5, in response to allergens. It is conceivable that before ED therapy, the patient's PBMCs, activated by food antigens, participated in the pathogenesis of EGE. T cell-mediated hypersensitivity to food antigens, measured by the lymphocyte proliferation test, decreased while he was receiving ED therapy. This decreased cell immune reactivity during ED may indicate a loss of circulating sensitized cells, suppression of the proliferative response related to the decreased antigenic exposure, or nutritional and metabolic changes associated with the ED therapy, such as essential fatty acid (EFA) deficiency-induced T-cell dysfunction. Besides completely removing food antigens from the intestinal lumen, ED therapy may also have a direct antiinflammatory effect by inducing EFA deficiency, thereby reducing eicosanoid production in the gut (23,24). However, subsequent supplementation and normalization of EFA status was not associated with a relapse of symptoms in our patient. Thus withdrawal of allergens was more likely to be effective than the EFA deficiency. Selenium deficiency, another nutritional complication of prolonged ED therapy encountered in this case, has been associated with peripheral and tissue eosinophilia in rats (25). However, our patient improved, and his eosinophil count normalized, despite selenium deficiency, suggesting that this micronutrient deficiency did not contribute to eosinophilia. We conclude that ED therapy represents a safe and an effective therapeutic option for severe EGE that is food allergen driven. In this case, the ED successfully reversed the enteropathy and growth failure in severe, steroid-dependent EGE. However, when used for several months, ED therapy may be complicated by micronutrient deficiencies such as those of selenium and essential fatty acids. Thus careful monitoring and supplementation is necessary if the formula used lacks adequate amounts of certain micronutrients. Acknowledgment: Supported in part by a grant from the Dairy Bureau of Canada (C.G., E.S.), a Chercheur-Boursier Clinicien Research Award from the FRSQ (E.S.), and a Fellowship Award from the Canadian Foundation for Ileitis and Colitis (C.J.). The authors thank Diane Lachapelle for assistance in preparing the manuscript.FIG. 1.: Growth curves in an adolescent boy with severe, steroid-dependent eosinophilic gastroenteropathy. Growth velocity corrected for bone age was well below the 3rd percentile, compatible with growth failure. Elemental diet (ED) therapy (arrow) resulted in an accelerated growth velocity and catch-up growth.