Elimination Diet or Swallowed Topical Steroid Treatment of Pediatric Eosinophilic Esophagitis: Five-Year Outcomes

Abstract

Background –

Eosinophilic esophagitis (EoE) is a chronic, antigen-mediated disease of the esophagus commonly treated with swallowed topical steroids (STS) or elimination diets (ED).Evidence of a long-term response to ED in pediatric patients is sparse.

Objective –

Our study sought to understand the natural history of pediatric EoE treated exclusively with ED and to examine a similar population of STS-treated EoE subjects. We hypothesized that long-term adherence to an effective ED would result in ongoing EoE disease remission.

Methods –

We conducted a retrospective study of pediatric EoE subjects who had at least two visits in a multi-disciplinary clinic. Subjects were identified who had a) a new referral with a suspected diagnosis of EoE; b) received either ED or STS alone, and; c) completed both a diagnostic and post-treatment endoscopy. Concomitant PPI use was allowed. We collected demographics, clinical features, treatment plans, and associated side effects on each subject. Remission was defined as <15 eosinophils/high-powered field.

Results –

We screened the electronic medical record from 2015–2016 for subjects cared for in the Gastrointestinal Eosinophilic Diseases Program (GEDP) who fit criteria for inclusion in this analysis. One hundred ninety-nine subjects were identified, 16 who received exclusive ED and 15 who were treated with STS. Treatment of these subjects was documented for 4.8 and 5.2 years respectively, p = 0.51. Significant differences between the groups were observed in average age at EoE diagnosis (3.5 years ED vs 7.8 years STS p = 0.002) and in number of endoscopies (6.6 in ED vs 4.5 in STS p = 0.03). Fifteen of 16 subjects treated with ED attained histologic remission. The initial effective ED removed a mean of 7.7 foods and the final ED removed a mean of 4 foods. No food impactions or esophageal dilations occurred in the ED group. The STS group required an average of 3.7 dose/formulation changes, 4 subjects required ≥1 dilation, 1 subject had 2 food impactions and 2 were diagnosed with adrenal insufficiency.

Conclusion –

Treatment with either ED or STS can lead to long-term remission of EoE. In this study, fewer side effects developed in the ED-group than the STS-group, but the validity of this conclusion is limited by the small sample size and reinforces the need for prospective study to explore these initial findings.

Introduction:

Eosinophilic esophagitis (EoE) is a chronic, primarily food antigen-mediated clinicopathologic condition of the esophagus with increasing incidence in both pediatric and adult populations^1,2. Swallowed topical steroids (STS) or elimination diets (ED) are common first-line therapies for EoE, and current guidelines recommend prolonged treatment³. A significant number of studies have investigated both induction and long-term maintenance of EoE remission with STS, but longitudinal outcomes for the use of ED in the literature are lacking^4–8. Advantages of ED compared to STS include potential identification of the underlying EoE allergen trigger, limited side effects when supervised by a dietitian, and possibly reduced cost⁹. When an effective ED is identified the expectation is long-term efficacy and minimal need for further dietary changes or procedures. Patients need only to avoid the foods that trigger their EoE, and in most this is 3 or fewer foods, although many triggers are dietary staples and decreased quality of life has been associated with various elimination diets^10–13. Other concerns regarding ED use are the initial requirement of more endoscopic evaluations while the diet is being developed, a negative impact on growth, lack of adherence, cost, and whether it is an appropriate therapy for patients with severe disease at the time of EoE diagnosis¹⁴. Short-term efficacy of ED has been shown, but robust data is lacking regarding the important issue of long-term outcomes, particularly in the pediatric population^{10, 15–18}. To address this information gap, the aim of this study was to document the longitudinal response of pediatric EoE subjects treated only with either ED or STS over approximately a 5-year period and to describe the outcomes and side effects associated with each therapeutic option.

Methods:

Study Design

We conducted a retrospective study of pediatric new regional patient visits for EoE between January 1, 2015 to December 31, 2016, cared for in the Gastrointestinal Eosinophilic Diseases Program (GEDP) at Children’s Hospital of Colorado and evaluated subjects treated with an elimination diet (ED). We also examined a similar group treated with swallowed topical steroids (STS) alone.

Over the evaluated time period, with the goal of 4–5 year patient follow-up, each new regional referral seen in the GEDP was assessed by a gastroenterologist and an allergist (MN and DA) to determine whether the visit was for an EoE diagnosis and whether initial treatment was limited to only STS or only initiation of an ED. For study inclusion, subjects needed to have had at least two visits to the GEDP and both a diagnostic esophagogastroduodenoscopy (EGD) as well as at least 1 follow-up EGD after initiation of therapy. Patients were excluded if they did not have an EoE diagnosis, if they were started on combination therapy with both ED and STS [concomitant proton pump inhibitor (PPI) use was allowed], if initial therapy was unclear based on chart review, or if they did not have a follow-up clinic visit or follow-up EGD, Figure 1. A detailed chart review of the EPIC electronic medical record was completed for each subject and a database was compiled with patient demographics, clinical information, gross endoscopic data, histologic data (peak eosinophil counts), and laboratory data. Clinical information included foods eliminated in an ED or prescribed STS (formulation and dose) and subsequent changes to the therapeutic plan. Foods eliminated from the diet were differentiated into those removed for EoE treatment, those avoided because of IgE-mediated food allergy, and those eliminated due to suspected food intolerance (example: lactose intolerance) or family preference/concerns about food intolerance. Foods eliminated as a therapeutic plan for EoE were agreed upon by a multidisciplinary team including an allergist, gastroenterologist, and registered dietitian. The majority of eliminated foods were removed empirically, but some subjects had foods removed as EoE therapy based on other factors such as skin prick testing. All dietary changes were verified by comparing physician clinic notes with registered dietitian documentation at each follow-up visit. The reasons for food elimination were confirmed by an allergist (DA). Symptoms of dysphagia, chest pain, vomiting, abdominal pain, reflux, behavioral changes, and other complaints were recorded based on reporting at the initial GEDP visit and changes in symptoms at follow-up visits. Growth data was obtained as BMI z-scores at diagnostic and last available EGD. Endoscopic Reference Score (EREFS) and peak eosinophil counts were obtained from all EGDs with histologic remission determined as <15 eosinophils/high-powered field (eos/HPF) and with strict histologic remission set at <6 eos/HPF¹⁹. Laboratory data included skin prick testing, serum food allergen-specific IgE testing (Immunocaps), nutritional surveillance labs (ferritin and vitamin D), and any fasting morning cortisol testing performed over the evaluation period. Fasting morning cortisol levels were only measured in subjects on STS to assess for adrenal insufficiency (AI). Skin prick testing to common food allergens was performed using DuoTip-Test^R II (Lincoln Diagnostics), histamine (positive) and saline (negative) controls and commercial food extracts (Stallergenes Greer). The mean diameters of wheal and erythema at 15 minutes were recorded. Skin tests with mean wheal diameters 3 mm larger than the negative control (saline) were considered positive. The study was approved by the local institutional review board.

Figure 1:

Study Flow Diagram showing detailed exclusion based on inclusion/exclusion criteria. This resulted in 15 subjects included in the STS-group and 16 subjects included in the ED-group.

Statistical analysis

Continuous variables were presented as mean plus range and/or standard deviation. Normally distributed data were compared using independent T-tests and comparisons within groups used paired T-tests. Non-normally distributed data were compared using KolmogorovSmirnov testing. Categorical variables were expressed as number and percentage within each category and were compared using Fisher’s exact testing. Findings were characterized as statistically significant with p-value < 0.05. Microsoft Excel Version 16.47.1 (Microsoft Corporation, Redmond, Wash.) and GraphPad Prism version 9.4.1 for macOS (GraphPad Software, San Diego, Calif.) were used for statistical analysis.

Results:

Of the 199 new regional referral patient visits over the study period, 16 subjects met criteria for inclusion into the ED group and an equivalent number of 15 subjects (from 2015 alone) met criteria for inclusion into the STS group. No differences were observed in PPI use between the groups 81% vs 87% treated; ED vs STS-group, p = 0.32. In the STS group, 7 of 13 subjects were treated with high-dose PPI (above 20 mg/day omeprazole equivalent), but only one subject started this high-dose PPI at or before EoE diagnosis. In the ED group, 6 of 13 subjects were treated with high-dose PPI and 5 of these subjects had their PPI started at or before EoE diagnosis. Eight of 15 (53%) subjects in the STS-group avoided foods due to IgE-mediated food allergy, parental suspicion of food intolerances, or celiac disease prior to the initiation of EoE treatment. The average subject age at EoE diagnosis in the ED group was significantly younger at 3.5 years (0.8 to 10.5 years) compared to 7.8 years (0.75 to 19.8 years) in the STS group, p = 0.002. Peak eosinophil count and EREFS scoring at time of EoE diagnosis were not different between the two treatment groups, Table 1.

Table 1:

Demographics, Clinical Characteristics, and Initial Endoscopic Findings

	Elimination diet (n = 16)	STS-treated (n = 15)	P-values
Age at EoE diagnosis; years (range)	3.5 (0.8 – 10.5)	7.8 (0.75 – 19.8)	0.002
Time under observation; years (range)	4.8 (1.1 – 7.8)	5.2 (2.4 – 8.5)	0.51
Sex; male%	81.3%	80%	0.99
Race; white%	93.7%	93.3%	0.99
Proton pump inhibitor use	81%	93.3%	0.32
Concomitant atopy
Food allergies	62.5%	66.7%	0.99
Allergic Rhinitis	56.3%	66.6%	0.72
Asthma	31.3%	53.3%	0.99
Eczema	31.3%	46.7%	0.47
Initial Endoscopy
Peak Eosinophil Count	51	55	0.49
EREFS Score	1.5	1.6	0.93

STS: swallowed topical steroids (fluticasone propionate, budesonide, ciclesonide) EREFS: Endoscopic Reference Score¹⁹.

Evaluation of practice patterns revealed that the initial effective ED eliminated an average of 7.7 foods and the final ED eliminated an average of 4 foods. ED subjects underwent an average of 6.8 EGDs in response to 6.3 dietary changes involving either food reintroduction or further eliminations, Table 2. At the initial GEDP clinic visit, all ED subjects reported between 1 and 4 presenting symptoms. Symptoms included reflux (63%), vomiting (50%), abdominal pain (50%), and dysphagia (44%). On follow-up, all subjects had improvement in at least 1 presenting symptom and only 2 subjects reported lack of complete symptom resolution, one with ongoing abdominal pain and another with continuing abdominal pain and vomiting. Overall, ED led to histologic remission (<15 eos/HPF) in 15 of 16 (94%) subjects. Using the stricter criterion for histologic remission of <6 eos/HPF, 12 of 16 (75%) met this threshold. Analyzing those subjects in our study who demonstrated improvement in eosinophilic inflammation (n = 15 of 16), the peak eosinophil count decreased significantly (50 eos/HPF vs 3.4 eos/HPF, untreated vs treated, p < 0.001), Figure 2. The longitudinal responses to ED are shown in Figure 3. Continued esophageal mucosal remission on a stable diet was observed in 84% (51/61) of EGDs performed. Milk was the most commonly identified food trigger and was excluded from the final diet in 13 of 16 (81%) subjects. After an effective ED was identified and instituted, 15 of 16 (94%) subjects tolerated reintroduction of foods over the study period, with 4 subjects (25%) on milk elimination alone (1 avoided all dairy, 3 avoided only unbaked dairy). No food impactions occurred in the ED group and no subjects required esophageal dilation. Finally, there was no difference in BMI z-scores when comparing the average z-score at the time of diagnostic EGD to final EGD, p = 0.26, Table 2.

Table 2:

Longitudinal Clinical and Endoscopic Outcomes

	Elimination diet (n = 16)			STS-Treated (n = 15)
Treatment of Changes; average	6.3; food removal/additions			3.7; STS formulation/type
Number of Foods Removed as EoE Treatment; average (range)	Max: 7.7 (1–10)	Final: 4 (1–9)		NA
Number of Endoscopies per patient; average (SD)	6.6 (3.2)			4.5 (1.8)
Total Number of Food impactions/esophageal dilations	0/0			2#/8^
Treatment Modality Change	3 (to STS)			2 (to ED)
Nutritional Supplementation	MVI: 11	Vit D: 4	Fe: 1	MVI: 11	Vit D: 5	Fe: 2
BMI Z-score; initial EGD/final EGD	−0.32		−0.54	0.31		0.45
Adrenal Insufficiency	0			2

^#One subject experienced both food impactions

^{^}Four total subjects underwent esophageal dilations

STS: swallowed topical steroids (fluticasone propionate, budesonide, ciclesonide), MVI: multivitamin, Vit D: vitamin D, Fe: ferrous sulfate.

ED: elimination diet

NA: not applicable

Figure 2:

Peak eosinophil counts in ED-treated subjects. Average peak eosinophil count in subjects with improvement in eosinophilic inflammation after effective ED initiation and at final EGD within the study period, n = 15 (*** = p <0.01).

Figure 3:

Longitudinal response to ED therapy. Change in peak eosinophil count in each EDtreated subject during sequential endoscopies. Subject 4 never responded to ED.

Within the STS group, 11 subjects were started on fluticasone and 4 were started on budesonide. Persistent symptoms led to an average of 3.7 treatment adjustments (either formulation or dose changes, supplementary Table E1) and each subject underwent an average of 4.5 endoscopies. At the initial GEDP clinic visit, all STS subjects reported between 1 and 5 presenting symptoms. Symptoms included dysphagia (60%), abdominal pain (60%), vomiting (53%), reflux (27%), chest pain (13%), and behavioral compliants (7%). Over the follow-up period, all except 2 subjects experienced complete resolution of all initially reported symptoms. One subject had no improvement in dysphagia and chest pain, and another improved in 3 of 5 initially reported symptoms, but had continuing dysphagia and feeding difficulties; however, neither of these patients had an esophageal dilation. From EoE diagnosis until the first follow-up EGD, peak eosinophil counts decreased signifcantly from 55 to 22 eos/HPF, p = 0.003, Figure 4. The longitudinal response to STS fluctuated widely over the evaluation period between subjects with esophageal mucosal remission observed in 48% of follow-up EGDs (25/52), supplementary Figure E1. Within this group, two food impactions occurred both in the same subject, and there were 8 dilation procedures performed with two subjects dilated once, one subject dilated twice, and one subject requiring 4 dilations, Table 2. No differences in BMI zscores were observed when comparing the average z-score at diagnosis to that at the final EGD, p = 0.57, Table 2.

Figure 4:

Peak eosinophil count in STS-treated subjects. Average peak eosinophil count from EoE diagnostic EGD compared to first follow-up after STS initiation (*** = p <0.01).

Five subjects changed treatments with 3 subjects transitioning from ED treatment to STS and 2 subjects transitioning from STS to ED. The reasons for stopping ED included improvement in weight gain with dairy reintroduction but with dairy identified as an EoE trigger, lack of histologic improvement, and difficulty adhering to the ED. Reasons for stopping STS included the development of iatrogenic adrenal insufficiency (AI) in one subject and family preference in another. The two subjects who developed AI were not symptomatic, but were identified through screening with fasting morning cortisol levels and diagnostic ACTH stimulation testing. One was prescribed concomitant inhaled steroids and intranasal steroids and the other intranasal steroids only. Both recovered, one after a change to ED with dairy elimination only and the other after a change to ciclesonide 320 μg, twice daily. Within the STS group, 7 subjects were treated with comcomitant inhaled steroids and 9 with intra-nasal steroids, supplementary Table E2. Comparison of the number of EGDs between the ED and STS groups was notable for the ED group undergoing 2.1 more EGDs on average, p = 0.03, Table 2. Fasting morning cortisol testing was not performed in the ED group, because no STS or systemic steroids were administered and no symptoms of AI were observed. The majority of subjects in both groups, 69% of ED and 73% of STS, Table 2, received multivitamin supplementation, two subjects in both groups had low vitamin D levels requiring supplementation, and iron supplementation was used by 1 subject in the ED group and 2 subjects in the STS group. No differences were observed in BMI z-scores at the time of EoE diagnosis or at final EGD between the ED and STS groups, but this analysis may have been limited by sample size.

Discussion:

EoE was first characterized in the 1990’s and differentiated from gastroesophageal reflux disease based on studies showing high rates of therapeutic efficacy using elemental diet to induce EoE disease remission^20–23. Lack of palatability of elemental formula, the inability of older children to orally ingest enough formula to meet nutritional requirements, difficulty in maintaining adherence, and the associated cost and social isolation resulted in using this therapeutic approach almost exclusively in young children. Due to these issues the majority of subsequent studies have focused on empiric elimination diets removing the most common allergenic foods such as dairy, wheat, eggs, soy, peanuts/tree nuts and fish/shellfish (six-food elimination diet; SFED), with dairy recognized as the most common food trigger. The short term efficacy in regard to histologic remission reported for the SFED of 74% is comparable to the reported 70–80% efficacy of using STS^{15, 24}. Other EDs have been investigated including the four-food elimination diet (milk/wheat/egg/soy; FFED), two-food elimination diet (milk/wheat), milk-only elimination, and the step-up elimination diet, and all have been shown to induce variable degrees of histologic remission over the short term^{14, 25, 26}. Eliminating foods based on skin prick testing or patch testing has proven to be less effective than the use of the FFED or SFED^{27, 28}. Commonly held concerns about instituting EDs include the requirement of more frequent endoscopies to accurately identify food triggers, quality of life issues such as social isolation associated with EDs, cost, growth concerns, and difficulties with patient adherence¹³. Additionally, long term efficacy has not yet been well documented, with studies generally following only a handful of subjects longitudinally for up to 3–4 years. The largest population with the longest follow-up period described to date is a retrospective survey study by Wang et al in 2018 that assessed factors influencing adherence to the SFED in 24 adult subjects who continued on SFED for an average of 3 years¹⁰. In this study, we report our experience over almost 5 years with 31 pediatric EoE subjects, 16 treated with long-term ED and 15 subjects treated only with STS.

Detailed investigation of these two treatment groups led to several notable findings. Although, the difference in the number of EGDs performed between the ED and STS groups reached statistical significance, whether the risk of two additional EGDs is clinically significant is debatable when there may be offsetting benefits to the ED treatment course. ED-treated subjects do not have an increased risk of AI compared to the general population, while our similar STS-treated population demonstrated 2 subjects who developed this side effect. Additionally, longitudinal histologic remission of subjects treated with effective ED (excluding positive food trigger re-introduction trials) appeared to be more consistent than in the STSgroup. No difference in peak eosinophil counts at EoE disease diagnosis between the groups and less fluctuation in the ED group over time than in the STS group hints that ED might lead to better disease control. Other explanations for this finding are different EoE disease phenotypes, age differential, or differences in treatment adherence between these groups²⁹. The possibility of more consistent disease control is notable in the context of a recent publication by Greuter et al 2020 that identified relapse of inflammation in 67% of a highly adherent adult population treated with STS⁷. If consistent application of ED truly maintains histologic remission more effectively, fewer endoscopies may be needed once innocuous foods are reintroduced and an effective ED is found. Another interesting finding from our study was that dilations were only required in the STS-treated group, which may have several underlying reasons . The most obvious explanation is that this was a result of the age differential between the two groups, allowing for disease progression over a longer time period prior to EoE diagnosis and implementation of therapy in the older STS group²⁹. It is also possible that the STS group had higher rates of a fibrostenotic EoE phenotype or more severe disease overall. Either of these explanations could lead to more advanced disease resulting in severe dysphagia or food impaction. Alternatively, it is possible that more dilations were needed because of poorer treatment adherence or a reduced therapeutic response in the STS group when compared with the ED group. Lack of differences in growth parameters in either group over the course of the study or between groups suggests no negative growth effects specific to either treatement modality. Alternatively, this lack of difference in BMI Z-score between ED and STS groups may be a result of small sample size. Finally, each group had subjects who were effectively transitioned to the other therapeutic approach (from ED to STS and vice versa). The reasons for these changes in therapy included the development of adrenal insufficiency (in one STS subject) and concerns about inadequate weight gain (in one ED subject), and were based on the notion that treatment is most effective when tailored to the preference of the patient and their family.

Limitations of this study include the retrospective design, which precludes reliable assessment of adherence and quality of life measurement, as well as the lack of standardization of the timing for endoscopies and follow-up visits. Additionally, there is inherent selection bias that may be present both from physician practice patterns regarding treatment allocation as well as in population acquisition by convenience sampling. However strengths of this study include detailed, objective data collected on each study subject and cross-referencing all documentation from different subspecialty providers to ensure the accuracy of each change in the therapeutic plan. Also, the high rates of histologic remission in the ED-treated group strongly supports that subjects were adherent to this form of therapy. Finally, while there is literature showing that EDs have notable impacts on quality of life, it is interesting that in this study similar numbers of patients in each group changed to the other treatment approach. This illustrates that the decisions leading to therapeutic changes involve consideration of multiple factors that vary over time due to therapeutic response, side effects of therapy, quality of life issues, social impact, and patient/family preference.

In summary, observation of pediatric EoE subjects treated only with either ED or STS over a 5-year period revealed sustained long-term marked reduction of symptoms, decreased peak eosinophil counts associated with continued histologic remission, and transition of a minority of subjects from one treatment course to the other. Each group, guided by symptoms and/or repeat endoscopic and histologic findings required changes to the initial ED or STS dose/formulation over time. These findings, combined with the significant age difference between the ED and STS groups support that either treatment option is reasonable and safe with high likelihood of a long-term favorable response depending on treatment adherence and the goals of the individual patient and their family. The significant difference in age between ED and STS groups raises the question of whether ED (similar to elemental diets) may be better tolerated in a younger age group. This may be due to increased parental control over diet in younger age groups. Additionally, parental concern about daily steroid dosing in a young child is often higher, and the social consequences of ED at a younger age may be less of an issue. Alternatively, STS treatment may be more acceptable initially in older children due to their concerns about dietary exclusions, social pressures, and some presenting with more advanced disease evidenced by fibrostenotic complications. Fortunately, those subjects (n = 5) that changed treatment modalities in our study did so without recurrence of severe inflammation. While the sample size is small, this lends support to the possibility of changing treatment modality if needed. Future prospective studies with larger numbers of subjects are needed to verify these findings and further explore differences between these treatment modalities. Outcomes from this work may aid to inform the future frequency of ED versus STS implementation.

Highlights Box:

1. What is already known about this topic:

   1. Elimination diets and swallowed topical steroids both induce histologic remission in EoE.
2. What does this article add to our knowledge:

   1. Children with EoE who are treated with ED or STS experienced long-term remission.
3. How does this study impact current management guidelines:

   1. Use of ED or STS can be effective long-term maintenance treatment for pediatric EoE.

Abbreviations:

EoE

eosinophilic esophagitis

STS

swallowed topical steroids

ED

elimination diet

EGD

esophagogastroduodenoscopy

GEDP

gastrointestional eosinophilic disease program

BMI

body mass index

eos

eosinophils

HPF

high-powered field

AI

adrenal insufficiency

FFED

four-food elimination diet

SFED

six-food elimination diet

PPI

protonpump inhibitor

EREFS

endoscopic reference score