Results
Patient Characteristics
The flow chart of the study population is shown in Supplementary Figure 1. Of 783 patients included into SEED, 323 were diagnosed by the senior author according to standardized protocols for assessment of clinical, endoscopic, histologic, and laboratory disease activity. Data on 123 patients were excluded due to either incomplete or missing information (records of 41, 37, 26, and 19 patients were missing data on the endoscopic features of EoE, symptom severity, laboratory workup, and histology, respectively). The remaining 200 EoE patients had complete datasets on clinical, endoscopic, histologic, and laboratory disease activity and were included for further analysis. The demographic and disease-specific characteristics of the patients included in this study are presented in Table 1. Of 200 patients, 153 patients were male (76.5%) with median age at diagnosis of 39 years (IQR, 28–50 years). Dysphagia was the leading symptom exhibited by 189 EoE patients (94.5%), followed by chest pain affecting 71 patients (35.5%). Allergies were identified in 132 patients (66.0%). An endoscopic bolus removal was performed in 56 (28.0%) patients; 28 patients underwent this procedure before, and 28 at the time of EoE diagnosis. Median diagnostic delay was 6 years (IQR, 2–12 years). Sixty patients (30.0%) had EoE symptom onset between 0 and 20 years of age, and 140 patients had symptom onset at older than 20 years of age. Twenty-two patients (11.0%) were diagnosed with EoE between 10 and 20 years of age. Diagnostic delay was longest in the young patient population (20 years of age or younger) and decreased with increasing age (Figure 1).
(Enlarge Image)
Supplementary Figure 1.
Patient flow.
(Enlarge Image)
Figure 1.
Diagnostic delay (time from first symptoms to diagnosis) in relation to age at EoE symptom onset (in decades). The horizontal line in the box represents the median; the box includes values from the 25 to 75 percentile. The figure shows that diagnostic delay is longest in patients with symptom onset 20 years of age or younger, and that diagnostic delay gradually decreases with increasing age of the patient.
Of 783 patients included into SEED, 460 EoE patients were diagnosed externally by other gastroenterologists in Switzerland. Due to the limited and nonstandardized data available for this group of patients, only data for the occurrence of strictures over time, which were available for 349 patients, were used in this article. Of these 349 patients, 277 (79.4%) were male, with median age at EoE diagnosis of 38 years (IQR, 26–53 years). Dysphagia was the predominant symptom in 336 patients (96.3%). Median diagnostic delay was 5 years (IQR, 2–10 years).
Endoscopic, Histologic, and Laboratory Characteristics at EoE Diagnosis
The key endoscopic, histologic, and laboratory findings at the time of EoE diagnosis are illustrated in Table 2. Strictures were found in 75 patients (37.5%). Further details on the number and other characteristics of esophageal strictures are provided in Table 3. Patients had a median of 1 stricture (IQR, 1–1; range, 1–3).
Features of active inflammation (edema, furrows, and whitish exudates) were found in 159 patients (79.5%) and features of fibrotic activity (strictures, rings, and crêpe-paper esophagus) were detected in 126 patients (63.0%) at EoE diagnosis. An endoscopic dilation of stricture(s) was performed in 27 patients (13.5%). No perforation was observed in dilated patients. Fifty-eight patients (29.0%) had peripheral blood eosinophilia (>0.35 g/L) and 124 patients (62.0%) had elevated IgE levels (>100 kU/L).
Evolution of Endoscopic Features Over Time
Analysis of the types of endoscopic features present at the time of EoE diagnosis stratified according to the length of diagnostic delay is shown in Figure 2. In this report, whitish exudates, furrows, and/or edema were considered as inflammatory features, and corrugated rings, strictures, and/or crêpe paper esophagus were considered as fibrotic features. Figure 2 shows that the prevalence of fibrotic endoscopic features at the time of EoE diagnosis either alone or in combination with inflammatory features increased with increasing duration of diagnostic delay from 46.5% (diagnostic delay 0–2 years) to 87.5% (diagnostic delay >20 years) (P = .020, Cochran-Armitage trend test). In addition, the prevalence of purely inflammatory features decreased with increasing duration of diagnostic delay period (P = .019). The prevalence of esophageal strictures positively correlated with the presence of endoscopic fibrotic features (Spearman's ρ = .3226; P < .001) and subepithelial fibrosis (ρ = 0.1927; P < .001).
(Enlarge Image)
Figure 2.
Types of the endoscopic features present at the time of EoE diagnosis stratified according to the length of diagnostic delay period. The frequency of patients exhibiting purely inflammatory endoscopic features at the time of EoE diagnosis decreased with the increasing length of diagnostic delay period, and the frequency of patients exhibiting a combination of inflammatory and fibrotic endoscopic features increased with increasing length of diagnostic delay period.
Prevalence of Strictures at EoE Diagnosis Significantly Correlates With Duration of Diagnostic Delay
We evaluated the prevalence of strictures at the time of EoE diagnosis stratified according to the length of the diagnostic delay periods. These findings are shown in Table 4. An increase in the stricture prevalence was observed over time, such that a group with a diagnostic delay of 0–2 years had a stricture prevalence of 17.2%, and a group with diagnostic delay of >20 years exhibited a stricture prevalence of 70.8%. The Cochran-Armitage trend test revealed a significant increase in the prevalence of strictures with increasing length of diagnostic delay period (P < .001). Figure 3 illustrates the "stricture-free" survival of the EoE patients during the diagnostic delay period encompassing >20 years. We were also interested if EoE in the group of 60 patients with symptom onset between 0 and 20 years of age progressed differently from the disease in the group of 140 patients with symptom onset at older than 20 years of age with respect to stricture formation. Therefore, we performed logistic regression modeling evaluating sex, presence of allergies, length of diagnostic delay period, presence of EoE family history, presence of blood eosinophilia, and elevated IgE levels as risk factors for the presence of strictures at the time of EoE diagnosis. These results are depicted in Supplementary Table 1. We found that the length of diagnostic delay was the only factor associated with the presence of strictures at the time of EoE diagnosis in the 20 years of age and younger symptom onset group (odds ratio [OR] = 1.116; 95% confidence interval (CI): 1.041–1.195; P = .002), as well as in the group of patients with EoE symptom onset at older than 20 years of age (OR = 1.088; 95% CI: 1.032–1.148; P = .002). We conclude that the population with symptom onset at 20 years of age or younger does not seem to be different from the group with EoE symptom onset at older than 20 years of age regarding stricture formation.
(Enlarge Image)
Figure 3.
Percentage of patients without strictures during the diagnostic delay period encompassing >20 years.
We also evaluated whether the relationship between the prevalence of strictures at the time of EoE diagnosis and the duration of untreated EoE can also be observed in the externally diagnosed EoE patients for which data on the diagnostic delay were available (349 of 460 patients [75.9%]). We observed the following stricture prevalence at the time of diagnosis for each corresponding diagnostic delay period: 18 of 155 patients (11.6%) for diagnostic delay period of 0–2 years; 8 of 60 patients (13.3%) for diagnostic delay period of >2–5 years; 6 of 30 patients (20.0%) for diagnostic delay period of >5–8 years; 13 of 43 patients (30.2%) for diagnostic delay period of >8–11 years; 4 of 15 patients (26.7%) for diagnostic delay period of >11–14 years; 6 of 11 patients (54.5%) for diagnostic delay period of >14–17 years; 8 of 19 patients (42.1%) for diagnostic delay period of >17–20 years; 9 of 16 patients (56.3%) for diagnostic delay period of >20 years. Using the Cochran-Armitage trend test, we found that the prevalence of strictures at the time of EoE diagnosis increased with increasing length of diagnostic delay period (P < .001) also in the externally diagnosed EoE patients.
Risk Factors for Stricture Formation
Using logistic regression modeling, we evaluated whether some patients are at higher risk than others to develop strictures. The dependent variables for the outcome "presence of stricture(s) at the time of EoE diagnosis" are described in the Methods section. The results of the logistic regression modeling are shown in Supplementary Table 2. Using univariate logistic modeling, we found that length of diagnostic delay (OR = 1.080 per year; 95% CI: 1.040–1.122; P < .001) was the only factor significantly associated with the presence of strictures at the time of EoE diagnosis. Of note, age at EoE symptom onset was not associated with the presence of strictures at the time of EoE diagnosis (OR = 0.988; 95% CI: 0.969–1.006; P = .195).
We hypothesized that persistent inflammatory activity could accelerate stricture formation in EoE patients. To evaluate a relationship between peak eosinophil counts in esophageal biopsies and stricture formation, we calculated the Spearman's correlation coefficient. However, the presence of strictures at the time of EoE diagnosis correlated neither with the proximal (Spearman's ρ = .048; P = .580) nor with the distal peak eosinophil counts (Spearman's ρ = .099; P = .248). Of note, tissue eosinophilia at a single time point, such as at the time of EoE diagnosis, is not likely to reflect the overall histologic disease activity over the entire course of the disease.