Food elimination in EoE: Milk before wheat, egg, and soy

To the Editor:

Eosinophilic esophagitis (EoE) is a chronic, allergic disease associated with symptoms of esophageal dysfunction and eosinophilic infiltration of the esophageal mucosa. Complete elimination of dietary antigens and supplementation with elemental formula results in histologic remission for approximately 90% of individuals, indicating that EoE is primarily a food-driven disease.¹ While this approach may be feasible in patients who rely exclusively on enteral tube feeding, it is not practical for most patients with EoE due to the risks of oral aversion, social isolation, cost, and the need for repeated endoscopies during food reintroduction. As a result, efforts to study less restrictive diet elimination approaches have emerged. Current guidelines for EoE management recommend these diets conditionally due to low or very low quality evidence.² Indeed, most studies examining diet elimination in EoE are single-arm, observational studies.³

In this issue of the Journal of Allergy and Clinical Immunology, Kliewer et al. report the results of a prospective, multicenter, randomized, comparative efficacy trial of 1-food (dairy) versus 4-food (dairy, wheat, egg, and soy) elimination diets in pediatric EoE.⁴ The authors hypothesized that a 4-food elimination diet (4FED) would result in a greater reduction in EoE symptoms than a 1-food elimination diet (1FED). Key secondary endpoints included the proportion of individuals achieving histologic remission [< 15 eosinophils per high power field (eos/hpf)], endoscopic severity, changes in the EoE transcriptome, and quality of life. Multiple validated outcome metrics were used to assess the primary and secondary endpoints including the Pediatric Eosinophilic Esophagitis Symptoms Score version 2.0 (PEESS v. 2.0), EoE Endoscopic Reference Score (EREFS), the EoE Histology Scoring System (EoEHSS), the EoE Diagnostic Panel (EDP), the Pediatric Quality of Life Inventory (PedsQL), Peds QL EoE Module, among others. The patients received standardized diet instruction from a dietician and diet and medication adherence were monitored through questionnaires. Sixty-three patients with active EoE non-responsive to proton pump inhibitor therapy were randomized to 1FED vs. 4FED for 12 weeks (Phase I). Esophagogastroduodenoscopies (EGDs) were performed at the end of each study phase to assess treatment response. 1FED nonresponders were assigned to treatment with 4FED, and 4FED nonresponders were assigned to treatment with swallowed glucocorticoids (SGCs) for 12 additional weeks (Phase II).

In Phase I, parent proxy symptoms scores were reduced more in the 4FED arm; however, child symptom and QoL scores were no different. Endoscopic, histologic, and transcriptomic findings were similar across treatment arms. Indeed, the rate of histologic remission was 44% for 1FED vs. 41% for 4FED. Withdrawal rates were higher for 4FED (32%) vs. 1FED (10.5%). Most patients withdrew before completing or initiating Phase II. Only 12.5% (1/8) of participants on 4FED and 50% (2/4) of participants on SGCs in Phase II achieved histologic remission. Interestingly, the authors identified peak tissue eosinophil counts as a predictor of treatment response to diet elimination. Milk specific IgG4 (sIgG4), the milk sIgG4/IgE ratio, and IgG4 to milk component protein Bos d5 were increased among responders to the 1FED.

This is an important study that rigorously evaluated EoE treatment outcomes after diet elimination across multiple domains. It provides compelling evidence that dietary therapy alone improves clinical symptoms and reverses disease features of pediatric EoE, including molecular changes in the EoE transcriptome. The results parallel the findings reported from the multi-center Six Food vs One Food Eosinophilic Esophagitis Elimination Diet (SOFEED) trial in adults that showed comparable efficacy between 1FED vs. 6FED (dairy, wheat, egg, soy, nuts, and seafood).⁴ In contrast to the SOFEED trial, subjects in this study eliminated foods from their diets for 12 instead of 6 weeks. Acknowledging age differences in study populations, the dropout rate was much higher among pediatric subjects on 4FED for 12 weeks (32%) compared to adult subjects on 6FED for 6 weeks (4.8%). This emphasizes the challenges associated with restrictive elimination diets and favors the relative simplicity and feasibility of dairy elimination as an initial approach to dietary management of EoE.

To date, there have been at least 4 prospective studies^4-7 examining dairy elimination for EoE treatment (Table 1). Most were conducted in children and demonstrated histologic response rates of 35.4-64.3% based on per protocol analysis. Even so, these studies may underestimate the overall efficacy of the 1FED as PPI responders were usually excluded. Moreover, it is not known whether extending treatment duration beyond 12 weeks may increase remission rates. Heterogeneity in symptom assessments preclude pooled comparisons of clinical efficacy across studies.

Table 1.

Prospective studies of 1-food (dairy) elimination diets for EoE treatment

First Author	Publication Year	Design	# Sites	Country	N 1FED	N per protocol	Study population (ages in yrs)	Diet duration (weeks)	N histologic responders to 1FED (%)†	Clinical response method
Kruszewski et al.5	2016	Prospective 2 arms 1FED vs. SGCs	1	USA	20*	14	Children/Adolescents (2–18)	6-8	9/14 (64.3%)	Peds QL, Peds QL EoE
Wechsler et al.6	2022	Prospective 1 arm 1FED	1	USA	54	41	Children/Adolescents (2–18)	8-12	21/41 (51.2%)	Peds QL, Peds QL EoE
Kliewer et al.7	2023	Prospective RCT 2 arms 1FED vs. 6FED	10	USA	67*	65	Adults (>18)	6	23/65 (35.4%)	EEsAI, EoE QoL-A, PROMIS
Kliewer et al.4	2024	Prospective RCT 2 arms 1FED vs. 4FED	10	USA	38*	34	Children/Adolescents (6-17)	12	15/34 (44.1%)	PEESS v. 2.0, Peds QL, Peds QL EoE, PROMIS

^*Subjects on 1FED a subset of larger study cohort

^†Histologic remission based on per protocol analysis

Abbreviations: 1FED – one-food elimination diet

An unresolved question is why dairy is the most important food trigger in EoE. This may be related to dose and nature of the allergen. Cow’s milk protein is one of the most common food antigens humans consume, particularly in early life. In addition, modern processing of cow’s milk with ultraheat treatment and high-pressure homogenization results in the formation of lipid-protein nanoparticles, which may facilitate antigen penetration of the esophageal mucosa and stimulation of immune responses.⁸

A key finding in this study was that peak tissue eosinophil counts predicted response to diet elimination. In addition, transition from 1FED to 4FED did not markedly improve remission rates, albeit a high dropout rate. Together, these observations suggest children and adolescents with more pronounced eosinophilic inflammation (peak eos/hpf > 42) are more likely to require SGCs or dupilumab to induce remission. Indeed, a “step up” approach where the diet is increasingly restricted⁹ may not yield favorable outcomes in those with severe disease. Notably, treatment with SGCs enhances the efficacy of subsequent PPI therapy in patients initially refractory to PPIs.¹⁰ Whether a similar approach enhances responsiveness to diet therapy in the future remains an open question. Positioning SGCs and dupilumab as induction agents for severe EoE and PPIs or a 1FED as maintenance therapy is an attractive treatment paradigm that could reduce side effects and cost. Future prospective studies of diet elimination should consider this approach.

A major limitation of this study was low recruitment resulting in a small sample size with imbalanced randomization. Unequal distribution of covariates may have biased the efficacy comparison between 1FED and 4FED and precluded detection of differences between groups. Congruent results in the SOFEED trial help mitigate this concern.⁷ A limitation of most diet studies in EoE is the inability to blind and control food antigen consumption. In both studies by Kliewer at al.^{4, 7}, diet adherence was excellent, highlighting the utility of education by a dietitian before pursuing diet elimination. However, the discrepancy between parent proxy and child symptom scores in the 4FED arm may reflect confirmation bias among parents anticipating treatment efficacy due to the greater stringency of the 4FED.

The studies by Kliewer et al and the Consortium of Eosinophilic Gastrointestinal disease Researchers (CEGIR) network^{4, 7} significantly advance our understanding of diet therapy in EoE. They demonstrate that a 1FED is comparable to more restrictive FEDs for EoE treatment and that dairy removal should be the first-line approach to empiric diet elimination. The degree of esophageal eosinophilia may influence which treatment is preferred to induce disease remission in pediatric patients. Additional studies are needed to establish why milk protein is the dominant food trigger of EoE and how FEDs are optimally situated in the EoE treatment paradigm.