Our analysis demonstrates that odynophagia and swallowing-independent retrosternal pain are, in addition to the well-known symptom of solid-food dysphagia, frequently occurring symptoms in EoE.

In 2008, Kapel et al. conducted a retrospective analysis of a national pathology database in the US, in which they identified 363 EoE-patients (321 adults and 42 children) from a cohort of upper endoscopies for any clinical reason [25]. All of these cases were further analyzed regarding the indication for upper endoscopy. In the adult population, the majority of patients were endoscopically assessed as a consequence to dysphagia (70.1%), followed by gastroesophageal reflux disease (GERD)/heartburn (27.1%). However, in 5.3% of the population the indication was odynophagia and in 3.4% chest pain. In children, on the other hand, the leading indication was GERD/heartburn (38.1%), followed by abdominal pain/dyspepsia (31%). Compared to the adult population, dysphagia was present in a considerably lower number of cases (26.2%). Analog to the adult population, odynophagia (2.4%) and chest pain (4.8%) were rare indications for upper endoscopy in children. In our cohort, consisting of only adult patients, odynophagia was present in almost one quarter (23.2%) of all patients. This rate is considerably higher than the one observed by Kapel et al. and almost as high as the rate they demonstrated for GERD/heartburn. Therefore, odynophagia as a key symptom in the diagnostic of EoE seems to be underestimated. In addition, the rate of retrosternal pain in our cohort (13.5%) was substantially higher than in the above-mentioned study (3.4%), as long as the terms of "chest pain" (used by Kapel et al.) and "retrosternal pain" (used in our study) are comparable. However, our results provide evidence that swallowing-independent retrosternal pain is a considerably frequent symptom in EoE-patients, comparable to the rates of abdominal pain/dyspepsia (13.1%) in adults and nausea/vomiting (14.3%) in children [25].

In accordance to the above mentioned, in their review Furuta et al. nicely pointed out, that adult patients most often describe dysphagia as their leading symptom, while children often complain of GERD-like symptoms, failure to thrive, emesis, and abdominal pain without dysphagia [26]. "Chest pain," the symptom most similar to retrosternal pain, was mentioned to be present in a range of 1–58% in adults and 17–20% in children, however, was not included in their recommendation of symptoms "suggestive of eosinophilic esophagitis" (Table 3 in their publication). Odynophagia, probably due to not existing data, was not mentioned at all.

The identification of symptoms potentially caused by an underlying EoE is especially important, as the diagnostic delay remains high (median = 4 years; in one-third ≥ 10 years [7]), resulting in the presence of pharmacologically not treatable esophageal strictures in > 30% of patients at diagnosis [4]. Part of the reason is the differing diagnostic approach to upper gastrointestinal symptoms between age groups, as endoscopic evaluation is performed less frequently in younger patients [7]. As EoE is a disease of the young, with more than half of the patients being younger than 31 years and 30% even younger than 21 years [7], the identification or EoE-typical symptoms is even more important.

In order to objectify disease activity several different scoring systems exist. In a comprehensive review, Warners et al. identified these and discussed their value and outcome measures [27]. In accordance to the above mentioned, the authors pointed out, that the existing scores mainly focus on dysphagia in the adult population, while in the pediatric population multiple other symptoms are included [27]. For example, the Dysphagia Symptom Questionnaire and the Straumann Dysphagia Index both focus on dysphagia, but are limited by neglecting other EoE-related Symptoms [28, 29]. In adults, Warners et al. conclude, that the Eosinophilic Esophagitis Activity Index for patient-reported outcome (EEsAI PRO) appropriately measures EoE symptoms in adults [8, 27]. In fact, the EEsAI PRO additionally includes questions regarding "pain when swallowing." In the underlying study, we were able to retract information regarding odynophagia ("pain when swallowing") from the physician questionnaire that is systemically used in the SEECS, which enabled the underlying analysis. In children, the Pediatric Eosinophilic Esophagitis Symptom Score (PEES v2.0) and a multiple-item symptom index developed by Aceves et al. include questions regarding "chest pain" and "painful swallowing" [30, 31]. However, as of now, neither odynophagia, nor retrosternal pain are included in diagnostic guidelines or follow-up scores as EoE-typical symptoms. As long as odynophagia and retrosternal pain are not officially recognized as potential EoE related, the diagnostic and clinical follow-up process is hampered in a potentially large group of patients.

In the present analysis, patients with odynophagia had a significantly higher symptom burden (judged by the physician and the patient), than patients without odynophagia. In addition, odynophagia occurred significantly more often in patients with higher endoscopic (EREFS score) and histologic activity (high number of eos per HPF and eosinophilic abscesses). The latter must be considered with reservations. We could show that overall histologic activity at one point was higher in patients with odynophagia. However, based on the available data we were not able to specifically link the symptom of odynophagia to active EoE as symptom assessment could have been taken place in then already treated disease.

Increased symptom severity seems logic. An association to higher endoscopic and histologic activity is questionable. Analog to our results, in the study of Kapel et al. patients with dysphagia had a significantly higher peak eosinophilic count [25]. However, the majority of existing literature could only show a moderate correlation between symptoms and endoscopic or histologic activity [32]. In addition, even with significant reduction of eosinophilic infiltration, clinical remission was not achieved in several studies in the past [33,34,35]. One reason might be, that the present study did not analyzes all features of the EoE histology scoring system (EoE-HSS), known to correlate better with EoE activity than the simple EoE count per HPF [36]. However, lamina propria fibrosis, one feature of the EoE-HSS, was evaluated and did not show a difference in patients with or without odynophagia. In addition, no association of current disease activity was shown in our study. One other histologic feature, that was not analyzed in the present study, was the degree of mast cell infiltration, which has been shown to be increased in EoE-patients [37, 38]. In regard to odynophagia, mast cells might play an important role, as results of a few recent studies suggest that they remain activated in disease remission with potential influence on persisting symptoms, particularly pain, without correlation with eosinophilic counts in otherwise controlled patients [39,40,41].

Matching the increased symptom severity, patients with odynophagia reported a significantly worse QoL, which again seems logic.

In accordance to patients with odynophagia, patients with retrosternal pain also had a significantly higher symptom severity (judged by physicians and patients). All other analyzed features, including QoL, showed no differences between patient with and without retrosternal pain, which is in line with the above discussed literature.

Of note, 5% of patients in the present study experienced both symptoms, odynophagia and swallowing-independent retrosternal pain, with a significant association (p = 0.004) to each other, indicating that pain in general is a feature of EoE. Of crucial importance regarding this observation, neither odynophagia, nor retrosternal pain was associated to concomitant erosive esophagitis, despite the fact that the most prevalent clinical symptom of erosive esophagitis is pain–with and without swallowing. Furthermore, only a few of the patients analyzed in our cohort underwent previous esophageal dilation. And even though the number of cases in our cohort were very low, only very few patients complaining about odynophagia or retrosternal pain were diagnosed with concomitant esophageal candidiasis or viral esophagitis. All of the above mentioned indicates that retrosternal pain (with or without swallowing) is a general feature of EoE and not a consequence to complications or treatment.

Data in the SEECS are being prospectively collected since 2016 according to a straight protocol. In addition, the sample size is rather large and data are gathered at multiple sites all over Switzerland. However, two major limitations hamper the results of the underlying study. First, as already discussed above, we do not have information regarding mast cell infiltration, which might play an important role in pain. Second, we are not able to discriminate patients with odynophagia and/or retrosternal pain with or without concomitant dysphagia. In other words, we cannot determine how many patients would have been additionally diagnosed, if odynophagia or retrosternal pain would have been the only upper GI symptom. For the same reason, it can only be speculated that the knowledge regarding odynophagia and retrosternal pain as potential EoE associated symptoms might influence the diagnostic delay. Also, as no stratification according to dysphagia severity and active disease was possible, the potential influence of concomitant dysphagia and active disease on symptom severity and quality of life is unclear. Lastly, as no data regarding outpatient reflux monitoring are available, it cannot be ruled out that some of the analyzed patients may have had concomitant non-erosive reflux disease causing or adding to odynophagia and/or retrosternal pain.

In conclusion, especially odynophagia, but also swallowing-independent retrosternal pain, are frequent features of eosinophilic esophagitis. Further studies are needed to confirm our findings, with special focus on concomitant dysphagia and mast cell infiltration in order to potentially implement both symptoms into clinical practice. Until then, EoE should be ruled out in any patient (in general and specifically the one with concomitant atopic disease) with an otherwise unexplained presence of the mentioned symptoms, independently of concomitant dysphagia.