Emerging incidence trends of eosinophilic esophagitis over 25 years: Results of a nationwide register‐based pathology cohort

Abstract Rationale Eosinophilic esophagitis (EoE) has emerged from a case‐reportable illness in the early 1990s to a distinct clinicopathological entity. Increasing worldwide incidences have been observed, although due to various study designs estimates are inconsistent. Aim To determine population‐based annual incidence rates over a time period of 25 years. Methods A nationwide register‐based pathology (PALGA) search was performed to identify reports describing esophageal eosinophilia between 1995 and 2019. EoE was identified if the diagnosis was confirmed by the pathologist. Crude incidence rates were estimated by the number of new EoE cases per year and matched with population data. Results Between 1995 and 2019, 7361 unique patients’ reports mentioned esophageal eosinophilia, of these 4061 were classified as EoE (71% male, mean age 37.9 ± 18.4 years). In total, 639 (16%) children (<18 years) were diagnosed. The EoE incidence increased from 0.01 in 1995 (95% CI: 0.0 – 0.04) to 3.16 (95% CI: 2.90 – 3.44) per 100.000 inhabitants in 2019. EoE was significantly more prevalent in males (OR 2.48 | 95% CI: 2.32 – 2.65; vs. females p < 0.001) and adults (OR 1.42 | 95% CI: 1.31 – 1.55; vs. children p < 0.001). Highest incidences were observed in 2019, being 4.37 (95% CI: 3.94 – 4.84) vs. 1.97 (95% CI: 1.68 – 2.29) per 100.000 males and females, respectively (p < 0.001). No seasonal variation was observed. Conclusion Over the past quarter century, the annual rates of newly diagnosed EoE patients raised dramatically and this increase has not reached a deceleration yet.


| INTRODUC TI ON
Eosinophilic esophagitis (EoE) is a chronic immune-mediated disease characterized by symptoms of esophageal dysfunction and infiltration of the mucosa with eosinophils. 1,2 Dysphagia and food impaction are the most typical complaints in adults, whereas gastroesophageal reflux disease (GERD)-symptoms and failure to thrive, feeding disorders, and abdominal pain-predominates in children. [1][2][3] EoE is diagnosed per consensus guideline if i) there are symptoms of esophageal dysfunction and ii) ≥15 eosinophils (eos) per high-power field (hpf) under routine light microscopy after hematoxylin and eosin staining are present in at least one esophageal biopsy. 4 Over the past years, EoE has emerged from a case-reportable illness in the early 1990s to a distinct clinicopathological entity. 5,6 Although clinicians are becoming more familiar with this relatively new disease, the expanded EoE frequency cannot be simply attributed to raised awareness alone and is outpacing any increase in diagnosis or detection. 7,8 The EoE epidemiology is rapidly evolving, and while genetic predisposition has been indicated, the increasing number of new EoE cases strongly suggests that (non-)allergic environmental disparities may also be critical in disease manifestation. 3,9 A worldwide tendency of rising EoE incidences has been reported, though current estimates are inconsistent due to variety in study designs (eg, register-based or insurance database vs. hospital-based case series), heterogeneous reporting, and modified diagnostic criteria. [10][11][12] Over successive years, the frequency of EoE in the Netherlands has also increased tremendously and it remains unclear whether this still continues to rise. 13,14 Hence, the Dutch register-based pathology database (PALGA) offers an unique opportunity to present an update of accurate EoE incidence rates with nationwide coverage. 11,12,15 Therefore, we aimed to 1) assess the annual EoE incidence rates within the entire population in the Netherlands over the past 25 years and 2) to identify demographic trends (ie, gender, age, and date of diagnosis) over time.

| Data collection
This cross-sectional study was conducted by using results from the nationwide network and registry from cyto-and histopathology in the Netherlands (PALGA). 15 This archive contains data from all 46 pathology laboratories and has national coverage since 1991.

| Diagnostic criteria and case-finding strategy
In this follow-up study, a similar diagnostic framework for case identification was used as was previously published by our research group. 13,14 In addition to the previous search , the national database PALGA completed a comprehensive search to retrieve all pathology reports, matching the terms "esophagus" in combination with "eosinophilic inflammation", "eosinophilic hyperplasia", "eosinophilia", "eosinophi", or "allerg" between January 1, 2016, and December 31, 2019. All reports including primary carcinomas or describing eosinophilia in other regions of the gastrointestinal (GI) tract were excluded. After the first search, all duplicates were removed. All patients that were included in one of our previous search strategies between the years 1995 and 2015, without confirmation of diagnosis, were re-reviewed and included if i) EoE was diagnosed based on a new pathology report and/or ii) EoE was suspected in retrospect based on previous reports and additional information with regard to the indication of performed esophagogastroduodenoscopy (EGD). 13,14 All cases were classified as EoE if 1) the diagnosis was confirmed by the pathologist and/or 2) the degree of esophageal eosinophilia in one biopsy sample (taken from ≥2 levels of the esophagus) was described as "markedly" (or words of comparable meaning), which was interpreted as similar to ≥15 eos/ hpf by the reviewers (BDvR, MJW, WEdR, MEB). All reports describing "mild" (ie, moderate or words of similar meaning) esophageal eosinophilia without mentioning a peak eosinophil count of ≥15 eos/ hpf were excluded. A manual review of all reports was performed by the first reviewer, and a second reviewer was asked in case of uncertainty to reach consensus. After the first manual review, additional information with regard to the indication of the performed EGD with biopsies was requested in case of uncertainty. A comprehensive evaluation of these reports including additional data was done by the reviewers, and cases were excluded if an EGD was performed due to suspicion of other potential causes of esophageal eosinophilia, such as drugs, parasitic infection, Cohn's disease, or GERD. Clinical information that was considered to further support the diagnosis of EoE included symptoms of dysphagia or food impaction, and typical endoscopic signs of EoE (eg, furrows, rings). Of note, in several cases the diagnosis of concomitant GERD was made if suggestive clinical signs were mentioned in the requested information with regard to the indication of the EGD performed. Endoscopic signs of reflux esophagitis or typical reflux-related symptoms were interpreted as being suggestive for the diagnosis or GERD. After re-review, all reports classified as EoE were included for final analysis. Furthermore, demographic data (gender, age, and date of diagnosis) and relevant histological features (spongiosis, micro-abscesses, basal zone hyperplasia, and sub-epithelial fibrosis) were derived from our database.

| Endoscopy with esophageal biopsy sampling
To estimate the number of EGDs with esophageal biopsy sampling performed between 1995 and 2019, the PALGA database was queried. Search criteria were "esophagus" and "biopsy." Of note, outcomes of this search yield an estimation of the total number of unique endoscopies with biopsies performed, considering that the search was not manually reviewed.

| Statistical analysis
Statistical analysis was performed by using IBM SPSS Statistics (version 25.0) (SPSS, Chicago, USA). To calculate the annual incidence rates between 1995 and 2019 the total Dutch population was considered to be at risk for developing EoE. Crude incidence values were calculated by using the total number of newly diagnosed EoE patients and matched with Dutch population data (https://www. cbs.nl). Incidence rates were calculated for the entire population and stratified for gender and age. Descriptive statistics were used to assess demographic characteristics, presented as mean (±SD) for normally distributed continuous data and percentages (%) for categorical data. Groups were compared with chi-square statistics and unpaired t test, as appropriate. A p-value of 0.05 was considered to be statistically significant. Odds ratio (OR) and 95% confidence intervals (CI) were calculated by using MedCalc Software Ltd (Ostend, Belgium).

| Case identification
The database search between January 1, 1995, and December 31, 2019, included a total of 14.963 reports, of which 5298 reports were excluded due to non-existing esophageal eosinophilia or F I G U R E 1 Flowchart of case identification. After the initial PALGA search, revisions, incorrect reports, and duplicates, as well as reports with absence of esophageal eosinophilia or presence of eosinophils in the gastrointestinal (GI) tract were excluded. A total of 7361 unique patients were eligible for review, of which 4061 cases were identified as eosinophilic esophagitis (EoE) in accordance with the conclusion of the pathologist   Figure 2.

| Patient characteristics
Furthermore, 108 patients were determined as EoE with a concomitant esophageal disease based on the conclusion of the pathologist. In total, 18 (17%) patients were identified with EoE and coexisting GERD, 64 (59%) patients with EoE and Barrett's esophagus, and 26 (24%) patients with EoE and esophageal candidiasis.
In addition, no seasonal variations in the diagnosis of EoE were observed within the entire study timeframe (Figure 3).

| Histological features
The 518 (13) 693 (17) 777 (19) 736 (18) 600 (15) 324 (8) 157 (4) Distribution of year of diagnosis in 5 years strata for male and female patients is presented in Figure 6.  was reduced by the consistent use of one similar diagnostic framework with nationwide coverage of histological data. 13,14 Regarding our diagnostic strategy, we consider these results to reflect a valid and consistent overview of EoE incidence rates over the past quarter century.

| Endoscopy with esophageal biopsy sampling
In conclusion, we present observations on escalating EoE incidences over a considerable time frame of 25 years in the Netherlands.
From these results, it is clear that EoE incidence has not stabilized yet and continues to rise. These findings underscore the need to further investigate the mechanisms underlying its pathogenesis and which dynamic environmental components could lead to such an expansion of EoE cases.