Empiric proton pump inhibitor therapy after esophageal food impaction may mask eosinophilic esophagitis diagnosis at follow-up (original) (raw)

Summary

Esophageal food impaction (EFI) is often the first presentation for patients with eosinophilic esophagitis (EoE); however, there is significant heterogeneity in the management of EFI. We aimed to study the impact of EFI management, particularly post-EFI medication prescriptions on EoE diagnosis, follow-up, and recurrence in patients with endoscopic features of EoE. In our retrospective study, adults presenting between 2007 and 2017 with EFI requiring endoscopic dis-impaction with endoscopic features of EoE (furrows, rings, and/or exudates) were included. We examined the impact of demographics and EFI management on EoE diagnosis, follow-up (esophagogastroduodenoscopy [EGD] or clinic visit within 6 months), and recurrence. We identified 164 cases of EFI due to suspected EoE. Biopsy was performed in 68 patients (41.5%), and 144 patients (87.8%) were placed on proton pump inhibitor (PPI) and/or swallow corticosteroids after EFI, including 88.5% of those not biopsied. PPI use at time of biopsy was negatively associated with EoE diagnosis (odds ratio: 0.39, confidence interval: 0.17–0.85). Sixty-one (37.4%) patients were lost to follow-up at 6 months. Recurrent EFI at 1 year occurred in 3.7% of patients. Medications, most commonly PPI, are frequently prescribed after EFI when the endoscopic features of EoE are present, which may mask the diagnosis of EoE on follow-up EGD. We estimated that for every five patients biopsied on PPI, one case of EoE is masked. As recurrent EFI within 1 year is uncommon, empiric therapy should be avoided until diagnostic biopsies are obtained. Further efforts to reduce loss to follow-up after EFI are also needed.

INTRODUCTION

Eosinophilic esophagitis (EoE), a chronic, immune/antigen-mediated disease with increasing incidence and prevalence, has become the leading cause of esophageal food impaction (EFI).1–3 EFI is a common presenting symptom for patients with EoE and presents a unique opportunity to diagnose and establish care for these patients.4 There are limited data and few guidelines on the management of EFI, particularly when EoE is suspected. The American Society for Gastrointestinal Endoscopy recommends obtaining biopsies at the time of EFI if EoE is suspected, but it does not address follow-up or prescribing practices.5

A survey of gastroenterologists identified significant heterogeneity in EFI management and found that 72% of pediatric and adult gastroenterologists do not routinely follow up with patients after EFI.6 Another study demonstrated that fewer than half of patients presenting emergently for EFI had adequate outpatient follow-up.7 Moreover, risk for loss to follow-up was greatest among those not biopsied or with ≥15 eosinophils per high-power field (eos/hpf) on biopsy, highlighting the missed opportunities to establish care for patients with EoE.

Importantly, EoE requires a histologic diagnosis of ≥15 eos/hpf in the absence of other causes of esophageal eosinophilia because endoscopic features have poor sensitivity.8,9 In retrospective studies, rates of esophageal biopsies at EFI range from 27 to 54% in patients presenting with EFI.3,10 Eighty-five percent of respondents to the aforementioned survey start patients on proton pump inhibitors (PPIs) after EFI, but only 46% of respondents wait for biopsy results before starting medical therapy.6 The authors estimated that over 10,000 patients at risk for EoE presenting to the ER with EFI are missed annually.

Approximately, 40% of patients with EoE will have a histologic response to PPI.11 The international consensus diagnostic criteria proceedings from the AGREE conference cautions that normal biopsies on PPI at initial endoscopy cannot rule out EoE because the biopsy results in the absence of PPI are unknown.8 Notably, there have been reports of PPI use masking both endoscopic and histologic features of EoE, and thereby, delaying diagnosis.12 Little is known about the impact of ubiquitous PPI prescribing after EFI on subsequent esophageal biopsies, particularly when EoE is suspected.

The primary purpose of this study was to investigate the impact of prescribing practices after EFI in patients with endoscopic features of EoE on the actual diagnosis of EoE at follow-up endoscopy. We hypothesized that high rates of empiric treatment with PPI or swallowed corticosteroids (SCs) at the time of EFI would result in lower rates of EoE diagnosis on follow-up esophagogastroduodenoscopy (EGD). We also investigated the impact of biopsy utilization, prescribing practices, and follow-up recommendations at the time of EFI on the actual follow-up and risk of recurrent EFI.

METHODS

Subjects

We conducted a retrospective cohort study of patients presenting emergently with EFI requiring EGD between January 2007 and December 2017 at a single tertiary care center. Adult patients (age ≥ 18) presenting with EFI who required EGD performed by a gastroenterologist for the relief of esophageal obstruction with endoscopic features of EoE as the cause of impaction were included. In patients presenting with >1 EFI in which EoE was the suspected etiology, only the first episode was included for analysis. Endoscopies performed by non-gastroenterology specialties were excluded. Patients with a prior diagnosis of EoE and patients presenting with non-edible ingestion resulting in a foreign body in the esophagus were also excluded.

Data collection

After obtaining Institutional Research Board approval, patients were identified using the International Classification of Diseases, 9th Revision Clinical Modification code ‘foreign body in the esophagus’ (935.1) or International Classification of Diseases, 10th Revision Clinical Modification codes ‘foreign body in esophagus’ (T18.1) and ‘food in esophagus’ (T18.12). Medical records were reviewed independently by authors L.H. and S.D. for appropriateness of inclusion in the study. Baseline patient demographics, including age, sex, ethnicity, and baseline medications, including PPI and SC, were collected. Endoscopic findings, interventions such as biopsy and dilation, medication prescriptions, and follow-up recommendations at the time of EFI were recorded. Patients were adjudicated according to the etiology of their EFI based on the findings and description of the endoscopic report. Patients with endoscopic features of EoE defined as linear furrows, esophageal rings, edema, and exudates as described by the performing endoscopic were included for further analysis. Because the EoE endoscopy reference score (EREFS) was introduced during the study period (2013) and was infrequently reported after its inception, EREFS was not included in our analysis.13

Follow-up and recurrence

Adequate follow-up was defined as an elective EGD or outpatient gastroenterology clinic appointment within 6 months of EFI. Patients with planned follow-up outside of our system were excluded from the follow-up analysis because the actual follow-up (clinic visit or endoscopy) could not be confirmed, nor data on the recurrent EFI data were available. Patient records were also reviewed to determine if the patients were ultimately diagnosed with EoE at any time during the study period. Recurrence was defined as any EFI requiring EGD after the date of initial EFI through 31 December 2019. Therefore, all patients had a minimum surveillance time for recurrent EFI or EoE diagnosis of 2 years.

Statistical analysis

Descriptive statistics were used to present the demographics, clinical characteristics, and outcomes of interest. Differences in continuous variables (i.e. age) were compared using a student _t_-test. Differences in categorical variables between the groups were tested using a Chi-Square test or Fisher exact test where applicable (expected cell sizes <5). We calculated the percentage of subjects receiving biopsies per calendar year for the cohort and used a linear regression to estimate the trends in biopsy rates per year. We used logistic regression to test the differences between levels with respect to the outcome for univariate measures and to model the outcomes in a multivariate manner, adjusting for many variables at once. Number needed to treat (NNT) was calculated from the odds ratios (ORs) and their confidence intervals were estimated in the logistic regression. Recurrence-free survival was calculated for the overall cohort in a time-to-event analysis. All analyses were performed using R (R Core Team, 2020) and were evaluated at the 0.05 significance level.

RESULTS

Our search strategy identified 1257 unique patient encounters of which 549 involved cases of EFI requiring EGD. Five cases were excluded because the procedure was performed by a non-gastroenterology specialist. Nine cases were excluded because the EFI endoscopy report was unavailable. The cause of EFI of the included 535 cases was adjudicated to the following etiologies: 55 established EoE, 170 endoscopic features of EoE, 161 esophageal stricture, 86 features suggestive of gastroesophageal reflux disease, including hiatal hernia or esophagitis, 26 suspected motility disorder, 11 esophageal cancer, 5 other, and 21 no description was provided. Of the 170 cases with features of endoscopic features of EoE, 6 subjects had multiple EFI and only their first EFI was included. We therefore included 164 individuals presenting with EFI with endoscopic features of EoE. Endoscopies were performed by 31 gastroenterologists with a median of four cases per provider (interquartile range (IQR): 2–6.5).

The majority of patients were male (81.1%) and Caucasian (95.1%) (Table 1). Baseline PPI before EFI use was recorded in 15 patients (9.1%). Esophageal rings (93.2%) were the most common endoscopic finding of EoE, followed by linear furrows (45.1%), mucosal edema (8.5%), and exudates (4.9%), with 70 patients (42.7%) having >1 endoscopic feature of EoE. Biopsy at the time of EFI was performed in only 68 patients (41.5%). There was no difference in the demographical characteristics between those patients biopsied and those not biopsied; however, patients with linear furrows (61.7 vs. 33%, P = <0.001) or >1 endoscopic feature of EoE (55.9 vs. 33%, P = 0.004) were more likely to be biopsied at the time of EFI. There was a non-significant trend toward increased biopsy rates over the study, with an increase of 1.7% per year (SE = 1.5, _R_2 = 0.126, P = 0.28). Dilation was seldom performed (2.4%) and only in those who were also biopsied (5.9 vs. 0%, P = 0.028). Medications were prescribed in 144 patients (87.8%) at the time of EFI and were almost exclusively PPI monotherapy (94.4%). PPIs were mostly prescribed at twice daily dosing (78.0%). Thirteen (12.6%) were prescribed once daily with low-dose PPI (20 mg), 10 (9.7%) once daily with high-dose PPI (40 mg), 28 (27.2%) twice daily with low-dose PPI, and 52 (50.4%) twice daily with high-dose PPI (PPI dosage was unavailable for 33, frequency for 5). For those not biopsied at EFI, 88.5% of these patients were placed on empiric medical therapy.

Table 1

Patient demographics categorized by biopsy status at EFI

EFI, esophageal food impaction; EGD, esophagogastroduodenoscopy; PPI, proton pump inhibitor; SC, swallowed corticosteroid; SD, standard deviation. Bold values indicate statistically significant findings – P-value <0.05.

Table 1

Patient demographics categorized by biopsy status at EFI

EFI, esophageal food impaction; EGD, esophagogastroduodenoscopy; PPI, proton pump inhibitor; SC, swallowed corticosteroid; SD, standard deviation. Bold values indicate statistically significant findings – P-value <0.05.

Biopsy results

Of the 68 patients biopsied at time of EFI, EoE was diagnosed in 51 (75.0%) (Fig. 1). Of the 96 patients who were not biopsied at the time of EFI, 61 (63.5%) returned for follow-up EGD within 6 months of whom 58 (95.1%) had esophageal biopsies. For those biopsied at follow-up, biopsies were diagnostic of EoE in 40 (68.9%), with a median follow-up time of 31.5 days (IQR: 18.3–50.8). There was no difference in time to follow-up EGD between those that had ≥15 eos/hpf and <15 eos/hpf on biopsy (39.0 ± 28.3 vs. 37.6 ± 33.7, P = 0.9).

Fig. 1

Flow diagram of esophageal food impaction (EFI) outcomes for patients with endoscopic features eosinophilic esophagitis (EoE). EGD, esophagogastroduodenoscopy; eos/hpf, eosinophils per high power field; GI, gastroenterologist; PPI, proton pump inhibitor; SC, swallowed corticosteroid.

The diagnostic rate of EoE for patients biopsied at follow-up was lower but was not significantly different than those biopsied at EFI (P = 0.550) (Table 2). Proximal and distal biopsies were obtained in 45% of patients at follow-up compared to 29.4% at the time of EFI (P = 0.067). Patients biopsied at follow-up were significantly more likely to be prescribed with PPI or SC at the time of biopsy (91.4 vs. 7.4%, P < 0.001). Univariate and multivariate analyses of potential factors predicting diagnosis of EoE are shown in (Table 3). Use of PPI at biopsy was negatively associated with a diagnostic biopsy of EoE on univariate (OR = 0.39) and multivariate (OR = 0.06) analysis after controlling for sex, time of biopsy, and biopsy location (two locations vs. one or unspecified). EoE diagnosis was not impacted by the frequency of PPI dosing (once daily: 56.3% vs. twice daily: 64.3%, P = 0.573) or PPI dose (high dose ≥40 mg/day and 57.1% vs. low dose: 59.4%, P = 0.591). Of note, all five patients who were not prescribed PPI at time of follow-up EGD had biopsies diagnostic for EoE. Age was also negatively associated with a diagnosis of EoE at biopsy (OR = 0.94). Two patients on PPI at follow-up with <15 eos/hpf on biopsy were later diagnosed with EoE; both had recurrent EFI prior to their diagnosis of EoE.

Table 2

Comparison of patients and biopsy outcomes based on time of biopsy

eos/hpf, eosinophils per high power field. Bold values indicate statistically significant findings – P-value <0.05.

Table 2

Comparison of patients and biopsy outcomes based on time of biopsy

eos/hpf, eosinophils per high power field. Bold values indicate statistically significant findings – P-value <0.05.

Table 3

Univariate and multivariate analysis to identify risk factors in diagnosing EoE (≥15 eos/hpf) at time of biopsy

CI, confidence interval; OR, odds ratio. Bold values indicate statistically significant findings – P-value <0.05.

Table 3

Univariate and multivariate analysis to identify risk factors in diagnosing EoE (≥15 eos/hpf) at time of biopsy

CI, confidence interval; OR, odds ratio. Bold values indicate statistically significant findings – P-value <0.05.

Based on the negative univariate association between medication and EoE diagnosis (OR = 0.39, Table 3), we estimated that for every 5.3 (95% confidence interval [CI]: 2.6–39.1) patients placed on empiric PPI therapy, one patient may have their EoE masked on follow-up endoscopy. In our cohort, of patients with <15 eos/hpf on follow-up EGD, five had between 5 and 14 eos/hpf on biopsy. All patients were prescribed with PPI at EFI, and one subject was later diagnosed with EoE. An additional patient prescribed with PPI at EFI had no intraepithelial eosinophils at their 6-week follow-up EGD but was diagnosed with EoE approximately 1.5 years after their initial EFI.

Follow-up and recurrence

An EGD or clinic visit after EFI was recommended for 156 patients (95.1%). When follow-up was recommended, patients who were not biopsied at EFI (n = 95) were only recommended to follow up with EGD; whereas, in patients who were biopsied, clinic follow-up was recommended for 24 (39.3%) and EGD for 37 (60.7%).

One patient was excluded from the follow-up data as they had planned follow-up outside of our health-care system. One-hundred and two patients (62.6%) had adequate follow-up within 6 months, most often with EGD (88.2%, n = 90) (Table 4). Thirty-four (35.4%) patients who did not have biopsies at the time of EFI were lost to follow-up. There was no difference in follow-up based on age, sex, prescription after EFI, or biopsy outcome at EFI. Follow-up rate at 6 months was greater among those diagnosed with EoE at EFI (OR = 2.06 [0.68–6.42], P = 0.2), albeit not significantly so. Patients recommended for follow-up at the time of EFI had somewhat higher rates of follow-up, but this also did not reach statistical significance (98.0 vs. 88.5%, P = 0.054).

Table 4

Comparison of patients with and without adequate 6-month follow-up

Table 4

Comparison of patients with and without adequate 6-month follow-up

Provider follow-up of the 68 EFI biopsy results was available for 61 cases. In the 47 cases when EoE was diagnosed on biopsy at EFI, PPI monotherapy was continued in 24, SC monotherapy was continued in 1, SC was started in 19 (only 1 provider explicitly instructed patient to discontinue PPI), and 3 providers referred patient to clinic to discuss therapy options. Of the 14 cases with <15 eos/hpf, 10 of the 11 patients started on PPI therapy were continued on PPI therapy after review of their biopsy results. One patient was switched from PPI to combination therapy with SC and another individual who was started on combination therapy after EFI was continued on that therapy due to high clinical suspicion for EoE in both cases despite biopsies showing <15 eos/hpf. Of the two subjects not started on therapy, one was asymptomatic at follow-up and remained off medications and the other was started on twice-daily PPI therapy for presumed GERD. Additionally, all 18 patients biopsied at follow-up with <15 eos/hpf were continued on PPI therapy at the time of the last gastroenterology follow-up. Only six of these patients established care in gastroenterology clinic after their follow-up EGD, of whom one patient was given a diagnosis of PPI-responsive EoE. One patient who was lost to follow-up after subsequent EGD later presented with recurrent EFI and was diagnosed with EoE 6 months after their recurrent EFI.

A total of 33 episodes of recurrent EFI occurred in 26 patients. There was no difference in the age (42.3 ± 10.4 vs. 40.7 ± 14.2, P = 0.600), sex (M: 88.5 vs. 79.7%, P = 0.416), or 6-month follow-up (61.5 vs. 62%, P = 1.00) between those with and without recurrent EFI. There was no difference in the recurrent EFI rates between those prescribed and those not prescribed PPI after EFI (15.0 vs. 16.0%, P = 0.911). Dilation at follow-up endoscopy was not protective against recurrent EFI (OR = 1.46, CI: 0.371–4.84, P = 0.6). Patients with recurrent EFI were more likely to be diagnosed with EoE at any time during the study period (88.5 vs. 57.2%, P = 0.002). Rate of recurrence at 6 months was 2.2%, 3.7% at 1 year, 7.9% at 2 years, and 15.9% at 10 years (Fig. 2).

Fig. 2

EFI recurrence in patients with endoscopic features EoE (minimum follow-up: 2 years).

DISCUSSION

EFI is often the first presentation for patients with EoE and offers a unique opportunity to confirm the diagnosis, begin disease-directed treatment, and establish care for these patients. Our study is the first to our knowledge to assess the impact of EFI management on EoE diagnosis and follow-up when EoE is initially suspected. In our cohort of 164 patients, we found low rates of esophageal biopsies at the time of EFI (41.5%). In the absence of a histologic diagnosis or pending biopsy, most patients (89%) were started on empiric therapy at the time of EFI.

While the diagnostic rate of EoE was not significantly lower at follow-up EGD compared to EFI, we found that PPI prescription was negatively associated with a diagnosis of EoE at time of biopsy, raising the concern that empiric therapy may mask a histologic diagnosis of EoE. We estimated that for approximately every five patients biopsied while on medical therapy, one case of EoE would be masked. Low-level esophageal eosinophilia may inappropriately be interpreted as gastroesophageal reflux when the patient in fact has EoE, highlighting the importance of interpreting available clinical, endoscopic, and histologic data appropriately. If PPI therapy is indicated for other reasons after EFI (i.e. erosive esophagitis and mucosal tear), the medication ideally should be stopped for several weeks before repeating diagnostic endoscopy. In our experience, endoscopy is typically arranged without further guidance from a gastroenterologist, and therefore, does not allow for a period of medication washout.

We also found that follow-up rates after EFI in patients with endoscopic features of EoE is poor: only 62.6% of patients had follow-up within 6 months of their EFI despite 95% of providers recommending some form of follow-up. While not all patients require in person post-EFI follow-up, especially if esophageal biopsies were adequately obtained, we found that over a third of patient not biopsied at the time of EFI never returned for esophageal biopsies despite EoE being their suspected diagnosis. Compared to Chang et al., our rate of follow-up was higher despite our more conservative definition; however, our study adds to their position on the need to improve follow-up after EFI.7 Recommendation of follow-up, which was common in our practice, appeared to improve the rates of follow-up among these patients. Failure to follow-up, particularly in those who have never been biopsied, may delay the time to diagnosis and appropriate medical and endoscopic therapy.

Lack of follow-up has also been associated with recurrent EFI.14 Overall, the incidence of recurrent EFI over our 11-year study period was 15.9%, with only 3.7% having recurrence within 1 year. Considering the aforementioned impact of medications on biopsy results and low rate of early recurrence, empiric therapy at EFI should be avoided if short-term follow-up endoscopy can be completed. Short-term follow-up within 6 months or dilation at follow-up endoscopy was not protective against recurrence, which is likely reflective of the chronic, inflammatory to fibrostenotic natural history of the disease. Dietary modifications, medical therapy, and endoscopic management likely impacted the frequency of recurrent EFI but could not be accounted for due to the nature of this study.

A major limitation of this study is the ability to thoroughly analyze the impact of adequate biopsies on EoE diagnosis. Current guidelines recommend multiple biopsies from at least two levels of the esophagus to ensure adequate tissue sampling and sensitivity to diagnosis EoE.8,15 Number of biopsies were unavailable for most patients and were therefore not included. Biopsy locations were not specified in the endoscopy or pathology report for 36.5% of patients, with nearly identical proportions of unspecified locations for those biopsied at EFI and follow-up. Despite this limitation, the rate of diagnosis of EoE in this population when biopsies were obtained was quite high (72.2%), which is consistent with the previously reported range for the positive predictive value (PPV) of endoscopic features for EoE (51–73%).9 Importantly, our study did not include patients without endoscopic features of EoE. Exclusion of these patients and the high prevalence of EoE in this enriched population likely explain the high sensitivity of endoscopic features observed in our study. It is important to note that approximately 17% of patients with EoE can have endoscopically normal mucosa, and some patients may simply present with an esophageal stricture. Excluding those with EFI due to malignancy or suspected/established EoE, we found that the rate of esophageal biopsy among other causes of EFI was only 6.7% (20/299), and of those, only three were diagnosed with EoE. Therefore, biopsy should also be strongly considered in all patients presenting with EFI, particularly among those with risk factors, including those with endoscopically normal mucosa.

The retrospective nature of the study only allowed for us to analyze the impact of medication prescription on biopsy results as opposed to actual medication use/adherence, which is estimated to be around 60% in patients on PPI.16 If non-adherence could be adjusted for, we may have observed an even greater effect of medication use on biopsy results. Inability to account for medication adherence may also explain why no association between PPI dosing or frequency and EoE diagnosis was identified as one may expect less of an impact of low-dose daily PPI compared to high-dose twice daily PPI. Additionally, we found on multivariate analysis, that after controlling for factors including medication prescription, there was a trend toward higher rates of EoE diagnosis at follow-up endoscopy compared to that at the time of EFI (OR = 8.7, P = 0.06). This trend may be explained by higher documented rates of proximal and distal biopsies at follow-up. However, it should be noted that PPI prescription is likely confounded with time of biopsy and biopsy location.

It is important to note that the study period occurred in an era in which consensus guidelines outlined PPI-responsive esophageal eosinophilia as a separate entity from EoE. In 2018, the AGREE consensus recognized PPIs as a treatment for EoE rather than a diagnostic test for esophageal eosinophilia.8 While PPI trials were uncommon for patients referred from community practices for outpatient management of EoE during this era, the explanation for the high rates of PPI prescriptions after EFI observed in our study is likely threefold: (i) the difference in study population—outpatient referral versus emergent presentation for EFI, (ii) greater recognition of EoE consensus guidelines at a tertiary care center, and (iii) indiscriminate PPI use after EFI, as 76.5% (420/559) of all EFI cases were prescribed PPI.17 While this major paradigm shift in the diagnosis and management of EoE occurred after our study period, we would not expect this to significantly impact the findings of this study as many patients and providers start with PPI therapy due to its ease of administration and proven safety profile.18 On the other hand, we found that 40.4% of providers started SC in addition to PPI after review of post-EFI biopsies; in light of growing recognition of EoE and experience in managing it with PPIs, we might anticipate lower rates of SC use in the post-AGREE era.

While endoscopists may plan for esophageal biopsies at the time of EFI, other circumstances, such as inadequate sedation, patient discomfort, and esophageal mucosal disruption, may ultimately preclude biopsy, something that could not be accounted for in this study. The biopsy rates reported in this study were comparable to those reported in previous studies. Another limitation of this study is that all data are from a single tertiary care center, which may lead to uniform styles of clinical practice; however, there likely was greater heterogeneity in practice style than expected for this study design as we report data across 31 gastroenterologists practicing in three different endoscopy centers within our tertiary care health system. Finally, our population was fairly homogenous, though this population represents those most likely to have EoE: Caucasian males. Nevertheless, we would not expect patient demographics to affect the endoscopic management of EFI, prescribing practices, or follow-up recommendations. Actual follow-up should be evaluated across more diverse populations due to well-established differences in access to health care based on race, ethnicity, and economic status.

In conclusion, our study furthers the narrative that management of EFI, particularly among those with endoscopic features of EoE, should be standardized with a focus on a high-quality esophageal examination, appropriate esophageal biopsies, and arranging adequate follow-up. Hiremath et al. have argued for a ‘best practice’ after EFI, which includes (i) esophageal biopsies at time of EFI, (ii) initiating PPI therapy in cases of suspected EoE, (iii) scheduling clinic follow-up after EFI, and (iv) scheduling follow-up EGD after EFI.6 Our data suggest that when biopsies cannot be obtained at EFI, empiric medications should be avoided due to the potential risk of masking EoE at follow-up endoscopy, especially as the risk of early recurrence is quite low. Lastly, future work should focus on developing systems to improve access to care and follow-up after EFI to reduce the high rates of loss to follow-up observed in our study.

Conflict of Interest

All authors have no conflict of interest to disclose.

Specific author contributions: Luke Hillman was responsible for the study concept and design, data collection, data interpretation, and manuscript drafting. Sarah Donohue was responsible for the data collection and manuscript drafting. Aimee Teo Broman was responsible for the data/statistical analysis and manuscript revision. Patrick Hoversten, Eric Gaumnitz, and Luis Lomeli were responsible for the study design/concept and manuscript revision.

Financial support: This work was supported by the Clinical and Translational Science Award (CTSA) program through the NIH National Center for Advancing Translational Sciences (NCATS) (grant UL1TR002373).

References

1

Desai

T K

,

Stecevic

V

,

Chang

C H

,

Goldstein

N S

,

Badizadegan

K

,

Furuta

G T

.

Association of eosinophilic inflammation with esophageal food impaction in adults

.

Gastrointest Endosc

2005

;

61

(

7

):

795

–

801

.

2

Dellon

E S

,

Hirano

I

.

Epidemiology and natural history of eosinophilic esophagitis

.

Gastroenterology

2018

;

154

(

2

):

319

–

32.e3

.

3

Hiremath

G S

,

Hameed

F

,

Pacheco

A

,

Olive

A

,

Davis

C M

,

Shulman

R J

.

Esophageal food impaction and eosinophilic esophagitis: a retrospective study, systematic review, and meta-analysis

.

Dig Dis Sci

2015

;

60

(

11

):

3181

–

93

.

4

Lucendo

A J

,

Molina-Infante

J

,

Arias

A

et al.

Guidelines on eosinophilic esophagitis: evidence-based statements and recommendations for diagnosis and management in children and adults

.

United European Gastroenterol J

2017

;

5

(

3

):

335

–

58

.

5

Committee ASoP

,

Ikenberry

S O

,

Jue

T L

et al.

Management of ingested foreign bodies and food impactions

.

Gastrointest Endosc

2011

;

73

(

6

):

1085

–

91

.

6

Hiremath

G

,

Vaezi

M F

,

Gupta

S K

,

Acra

S

,

Dellon

E S

.

Management of esophageal food impaction varies among gastroenterologists and affects identification of eosinophilic esophagitis

.

Dig Dis Sci

2018

;

63

(

6

):

1428

–

37

.

7

Chang

J W

,

Olson

S

,

Kim

J Y

et al.

Loss to follow-up after food impaction among patients with and without eosinophilic esophagitis

.

Dis Esophagus

2019

;

32

(

12

):1–4.

8

Dellon

E S

,

Liacouras

C A

,

Molina-Infante

J

et al.

Updated international consensus diagnostic criteria for eosinophilic esophagitis: proceedings of the AGREE conference

.

Gastroenterology

2018

;

155

(

4

):

1022

–

33.e10

.

9

Kim

H P

,

Vance

R B

,

Shaheen

N J

,

Dellon

E S

.

The prevalence and diagnostic utility of endoscopic features of eosinophilic esophagitis: a meta-analysis

.

Clin Gastroenterol Hepatol

2012

;

10

(

9

):

988

–

96.e5

.

10

Sperry

S L

,

Crockett

S D

,

Miller

C B

,

Shaheen

N J

,

Dellon

E S

.

Esophageal foreign-body impactions: epidemiology, time trends, and the impact of the increasing prevalence of eosinophilic esophagitis

.

Gastrointest Endosc

2011

;

74

(

5

):

985

–

91

.

11

Hirano

I

,

Chan

E S

,

Rank

M A

et al.

AGA institute and the joint task force on allergy-immunology practice parameters clinical guidelines for the management of eosinophilic esophagitis

.

Gastroenterology

2020

;

158

(

6

):

1776

–

86

.

12

Odiase

E

,

Schwartz

A

,

Souza

R F

,

Martin

J

,

Konda

V

,

Spechler

S J

.

New eosinophilic esophagitis concepts call for change in proton pump inhibitor management before diagnostic endoscopy

.

Gastroenterology

2018

;

154

(

5

):

1217

–

21.e3

.

13

Hirano

I

,

Moy

N

,

Heckman

M G

,

Thomas

C S

,

Gonsalves

N

,

Achem

S R

.

Endoscopic assessment of the oesophageal features of eosinophilic oesophagitis: validation of a novel classification and grading system

.

Gut

2013

;

62

(

4

):

489

–

95

.

14

Prasad

G A

,

Reddy

J G

,

Boyd-Enders

F T

,

Schmoll

J A

,

Lewis

J T

,

Wongkeesong

L M

.

Predictors of recurrent esophageal food impaction: a case-control study

.

J Clin Gastroenterol

2008

;

42

(

7

):

771

–

5

.

15

Dellon

E S

,

Gonsalves

N

,

Hirano

I

et al.

ACG clinical guideline: evidenced based approach to the diagnosis and management of esophageal eosinophilia and eosinophilic esophagitis (EoE)

.

Am J Gastroenterol

2013

;

108

(

5

):

679

–

92

quiz 93

.

16

Hungin

A P

,

Hill

C

,

Molloy-Bland

M

,

Raghunath

A

.

Systematic review: patterns of proton pump inhibitor use and adherence in gastroesophageal reflux disease

.

Clin Gastroenterol Hepatol

2012

;

10

(

2

):

109

–

16

.

17

Whitson

M J

,

Lynch

K L

,

Yang

Y X

,

Metz

D C

,

Falk

G W

.

Lack of proton pump inhibitor trial prior to commencing therapy for eosinophilic esophagitis is common in the community

.

Dis Esophagus

2018

;

31

(

5

):1–7.

18

Tourlamain

G

,

Garcia-Puig

R

,

Gutierrez-Junquera

C

et al.

Differences in management of eosinophilic esophagitis in Europe: an assessment of current practice

.

J Pediatr Gastroenterol Nutr

2020

;

71

(

1

):

83

–

90

.

© The Author(s) 2021. Published by Oxford University Press on behalf of International Society for Diseases of the Esophagus. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com