This systematic review has been registered in the PROSPERO International prospective register of systematic reviews (www.crd.york.ac.uk/PROSPERO; register no. CRD42014006981), and was reported in accordance with the PRISMA statements.
Selection of Studies
A systematic literature search was performed independently by two researchers (AA and JMT) in three major bibliographical databases (PUBMED, EMBASE and Scopus) for the period up to December 2013. The search was not restricted with regard to date or language of publication. To this end, a predetermined protocol was used in accordance with the quality of reporting meta-analyses of observational studies in epidemiology.
Comprehensive search criteria were used to identify articles dealing with the relationship between EoE and CD. We consulted the thesauri for MEDLINE (MESH) and EMBASE (EMTREE) using the following search strategy: eosinophilic AND ('esophagitis'/exp OR esophagitis) OR eosinophilic AND ('oesophagitis'/exp OR oesophagitis) AND (coeliac) AND ('disease'/exp OR disease) OR coeliac AND ('disease'/exp OR disease) OR 'gluten'/exp OR gluten AND ('enteropathy'/exp OR enteropathy) OR 'hla'/exp OR hla) AND ('antigens'/exp OR antigens).
For the Scopus database, only free text searches with truncations were carried out. The search was not restricted with regard to date or language of publication.
We also examined the reference lists from retrieved articles and abstracts of conference proceedings (abstract books of the annual Digestive Diseases Week, American College of Gastroenterology Meeting, and the United European Gastroenterology Week for the period between 2004 and 2013) to identify additional relevant studies. Two reviewers (AJL & AA) independently screened the database search for titles and abstracts. If any of the reviewers felt that a title or abstract met the study eligibility criteria, the full text of the study was retrieved.
Studies were included in the systematic review if they provided original data on the concomitant diagnosis of CD in individual patients or patient series with EoE, or, alternatively, if they described a diagnosis of EoE in individuals or series of patients with CD. Such studies were included irrespective of study design (i.e., randomised controlled trials [RCT], observational prospective and retrospective studies, and case series reports).
Studies evaluating a gluten-free diet-based intervention in EoE patients with CD were also considered if objective quantitative data on efficacy in terms of histological response were provided. EoE remission was considered to be a peak eosinophil count <15 eos/high power field (hpf) in oesophageal biopsies after a GFD-based dietary treatment.
The following were excluded from our analysis:
Reviews on the treatment of EoE that did not provide original data on dietary therapy, along with clinical guidelines and consensus documents.
Studies not carried out on humans.
Studies providing duplicated information (i.e., repeated abstracts presented at different congresses or abstracts published later as a full paper).
Subsets of cases or controls from a previously published article by the same authors.
Studies using a gluten-free diet intervention simultaneously with another therapeutic alternative capable of reducing oesophageal inflammation (topical and systemic steroids and/or immunomodulatory drugs) were not considered in the evaluation of the efficacy of a GFD, but were included in prevalence analyses.
Cohort studies, case series and case reports were evaluated for quality only if the article described all patients' demographical data, diagnostic criteria for EoE and CD, the proportion of EoE patients among CD patients (or vice versa) and study design. The effects of a gluten-free diet on resolution of EoE were also assessed. Quality assessment was checked with a specific evaluation form for observational studies developed by our group and based on the Strobe statements. A study was considered to be at low risk for bias if each of the bias items could be categorised as low risk. On the contrary, studies were judged to have a high risk of bias if even one of the items was deemed high risk. Two investigators (AJL & AA) independently gave each eligible study an overall rating of high, low or unclear risk of bias; if disagreements arose, a third reviewer (JMT) was consulted.
Two reviewers (AJL, AA) independently extracted relevant information from each eligible study using a standardised data extraction sheet and then proceeded to cross-check the results. The data extracted included the trial study areas, the last name of the first author, year of publication, age and gender of study participants, sample size, methodological design, study period, and, whenever possible, the effectiveness of a GFD on EoE. At the same time, data on the key outcomes, including prevalence of EoE among CD patients and/or prevalence of CD among EoE patients, were extracted from all included studies. Disagreements between reviewers regarding data extraction were resolved through discussion.
Response percentages for dietary intervention were summarised with the aid of a fixed- or random-effects meta-analysis weighted for inverse variance following DerSimonian and Laird's method. Summary estimates, including 95% confidence intervals (CI), were calculated for the prevalence of EoE among CD and vice versa, as well as for the efficacy of a GFD on EoE remission.
Heterogeneity between studies was assessed by means of a chi-square test (Cochran Q statistic) and quantified with the I2 statistic. Generally, I2 was used to evaluate the level of heterogeneity, assigning the categories low, moderate and high to I2 values of 25%, 50% and 75% respectively. Publication bias was evaluated with the aid of a funnel plot, the asymmetry of which was assessed with Begg–Mazumda's rank test along with the Egger and Harbord tests.
For the primary outcome, planned subgroup analyses were performed based on the primary population studied (patients with EoE or patients with CD) and age (adults vs. children).
A sensitivity analysis was performed with regard to quality (risk of bias) and type of document (full-length article vs. abstract presented at conference proceedings). All calculations were made with StatsDirect statistical software version 2.7.9 (StatsDirect Ltd, Cheshire, UK).
Aliment Pharmacol Ther. 2014;40(5):422-434. © 2014 Blackwell Publishing