Efficacy of Proton Pump Inhibitor Therapy for Eosinophilic Oesophagitis in 630 Patients

Results From the EoE Connect Registry

Emilio J. Laserna-Mendieta; Sergio Casabona; Danila Guagnozzi; Edoardo Savarino; Antonia Perelló; Antonio Guardiola-Arévalo; Jesús Barrio; Isabel Pérez-Martínez; Anne Lund Krarup; Javier Alcedo; Susana de la Riva; Esther Rey-Iborra; Cecilio Santander; Ángel Arias; Alfredo J. Lucendo


Aliment Pharmacol Ther. 2020;52(5):798-807. 

In This Article

Patients and Methods

Study Protocol

This cross-sectional analysis focused on the "European Registry of Clinical, Environmental and Genetic Determinants in Eosinophilic Oesophagitis" (EoE CONNECT), an international multicentre prospective-maintained non-interventional registry that was started in 2016 and was promoted by United European Gastroenterology as a part of the Link Award program "Harmonizing diagnosis and therapy of Eosinophilic Oesophagitis across Europe (HaEoE-EU)". EoE CONNECT is managed by EUREOS, the European Society of EoE.[22]

Patients of all ages with a confirmed diagnosis of EoE, no previous PPI treatment but having received at least one therapeutic intervention based on a PPI drug as first- or second-line treatment and who had provided informed consent to be registered on the EoE CONNECT database were included in the present study. Prospective treatment data were registered sequentially, and new sequences were created each time a different treatment (active principle, formulation or dose) was administered to a patient. The EoE CONNECT registry was approved by Research Ethics Committees in all participating centres. All co-authors had access to the study data and reviewed and approved the final manuscript.

Data Collection

Information is imputed onto EoE CONNECT by practitioners during face-to-face clinical appointments. Variables retrieved for this study included patients' demography, EoE characteristics at diagnosis (phenotype, dysphagia symptoms score and endoscopic findings), starting date of PPI therapy used for EoE (active principle, dose regimen and daily dose), clinical and histological response to therapy and evaluation date for PPI therapy. Endoscopic findings at baseline endoscopy were assessed by the EREFS scoring system;[24] rings and strictures were classified as fibrotic findings, while oedema, furrows and exudates were defined as inflammatory features.[25]

Monitoring and Quality Data

The database was monitored and individual treatment data were manually revised to evaluate whether the study selection criteria were met, the information was correctly registered and ultimately, to ensure the correct order of therapies to guarantee the highest scientific and ethical standards. Data completion was assessed based on the following three pivotal group of variables: "baseline characteristics," "PPI treatment," and "effectiveness of results". Duplicates were removed; data discordances were resolved by querying the investigators and through group e-mailing. Additionally, after data extraction and prior to statistical analysis, the database was reviewed for inconsistencies and subsequently subjected to data cleaning.


Active Disease. Active disease was defined as a peak eosinophilic infiltrate by ≥15 cells per high power field (hpf) at any oesophageal level together with ≥5 points in the Dysphagia Symptoms Score, a nonvalidated measure instrument previously used in trials assessing drugs[26,27] and diets[28,29] involving adolescent and adult EoE patients. A Dysphagia Symptoms Score ≥8 was considered as severe dysphagia as previously described.[22] Subjective symptom intensity reported by either children or parents was considered for younger children.

PPI Doses. Standard doses of PPI included omeprazole 20 mg, pantoprazole 40mg, esomeprazole 20 mg, lansoprazole 30 mg and rabeprazole 20 mg daily, following the proposal of The World Health Organization Collaborating Centre for Drug Statistics Methodology regarding treatment of gastro-oesophageal reflux disease (http://www.whocc.no/atc_ddd_index/?code1/4A02BC&showdescription1/4yes, accessed April 4, 2020), consensus guidelines[30] and experimental research.[21–33] Double doses or higher of the above were considered high-dose PPI,[34] and a low dose was defined when PPIs were given under standard or half-standard doses.

Evaluation of Response. Deep histological remission was defined as an eosinophil peak count of <5 eosinophils/hpf at all oesophageal levels after therapy; histological remission was considered as a peak count between 5 and 15 eosinophils/hpf.

Symptomatic improvement was independently assessed by changes in Dysphagia Symptoms Score reported by patients and by clinicians' perceptions. A decrease of more than 50% in baseline Dysphagia Symptoms Score after therapy was considered clinical remission in older children and adults, as previously defined;[22,28,29] a symptomatic improvement ≤50% from baseline was considered as clinical response. For younger children, any subjective improvement in symptoms reported by either children or parents was considered as clinical remission. In addition, clinicians semi-quantitatively scored changes in symptoms from the initiation of therapy as complete clinical remission, partial remission or no response.

Clinico-histological remission was defined as the simultaneous combination of symptomatic remission or improvement and all degrees of histological remission (peak eosinophil count <15 eosinophils per hpf) in the same patient after therapy.

Lack of efficacy was defined either as maintenance or worsening of patient's symptoms combined with persistence of histological activity of the disease at the end of PPI therapy, or a situation that led the physician to escalate the dose of PPI or change to an alternative drug or diet.

Statistical Analysis

Means and SDs, or alternatively medians and interquartile ranges (IQR), were reported for continuous variables and proportions for categorical data. Frequency tables were generated for treatment use and effectiveness. Contingency tables to assess demographical and clinical factors influencing treatment response rates were produced and analysed by chi-square or Fisher exact (univariate) test. A binary logistic multivariate regression analysis was performed to assess the overall effect of PPI treatment over variables identified in univariate analyses. All analyses were carried out using PASW 18.0 statistical analysis software (SPSS Inc). Statistical significance was considered when P < 0.05. Odd ratio (OR) was reported for those variables reaching statistical significance.