Biologics for eosinophilic oesophagitis: a systematic review and meta-analysis

1. Introduction

Eosinophilic gastrointestinal diseases (EGIDs) constitute a collection of gastrointestinal disorders distinguished by tissue eosinophilia and the manifestation of symptoms indicative of oesophageal or gastrointestinal dysfunction. According to the location of the eosinophilic infiltrate, EGIDs are classified as eosinophilic oesophagitis (EoE), gastritis (EoG), enterocolitis (EoN), and colitis (EoC) [1,2]. Clinical symptoms of EGIDs vary depending on the disease location, extent, and depth in the gastrointestinal system [3]. EoE, as the most common and extensively studied type among all EGIDs, boasts a prevalence of 42.2 cases per 100,000 adults and 34 cases per 100,000 children [4,5]. Patients suffering from EoE may present with heartburn/chest pain, dysphagia, nausea, or food impaction [6,7]. Unlike EoE, the symptoms of non-EoE EGIDs are much less specific and difficult to distinguish from other GI disorders [8]. While non-EoE EGIDs are far less common, a population-based study in the US found that the prevalence of EoG plus EoN was 5.1 per 100,000 people, and the prevalence of EoC was only 2.1 per 100,000 people [9]. A significant proportion of patients with EGIDs will have multiple sites of involvement [10]. However, the prevalence of EoE and other non-EoE EGIDs might be underestimated due to poor awareness and diagnostic delays [11].

The pathophysiological mechanisms of the various types of EGIDs are not completely known. The type 2 (T2) inflammatory response seen in EoE is similar to that seen in atopic illnesses like asthma, where elevated amounts of the characteristic T2 cytokines, IL-4, IL-5, and IL-13, are released [12]. These cytokines lead to an increase in eosinophil count, which may result in movement disorders and abnormalities in the esophageal epithelial barrier [13,14]. In contrast, non-EoE EGIDs are rarer than EoE and thus less well understood [15]. A similar T2 signature in EoE has been identified in EoG and EoN, but whether EoE and all non-EoE EGIDs share a common etiology is unclear [16].

The most prevalent therapies in EoE include pharmacologic measures, diet modification, and oesophageal dilatation for complications [17]. However, numerous patients with EoE did not attain histological remission despite the administration of corticosteroids or proton-pump inhibitors, and pharmacologic or dietary therapies have no proven benefit in non-EoE EGIDs as well [18]. With the expanding understanding of the pathogenesis of EoE and other non-EoE EGIDs, more and more biologic modulators were explored on them which exert therapeutic effects by targeting potential pathways [12]. Biologics that target Type 2 cytokines such as IL-5 (mepolizumab, reslizumab), IL-4/IL-13 (dupilumab), or IL-13 (dactrekumab, cendakimab) and that target Sialic acid-binding immunoglobulin-like lectin 8 (Siglec-8) receptor on eosinophils/mast cells have shown clinical symptom relief or histologic improvement in patients with EGIDs [19,20]. The therapeutic efficacy of biologics for EoE and other non-EoE EGIDs is still under debate. Therefore, we aim to assess the efficacy and safety of biologics compared to a placebo in the treatment of EoE patients by performing a comprehensive meta-analysis of existing high-quality, double-blind randomized controlled trials (RCTs).

2. Methods

PRISMA guidelines were followed in performing this meta-analysis (PRISMA Checklist is shown in Table S1), which was submitted to the International Prospective Register of Systematic Review (PROSPERO: No. CRD42023440744) [21].

Two reviewers independently searched the four databases of PubMed, Embase, Web of Science, and Cochrane Library for relevant literature. We also searched ClinicalTrials.gov and the European Medicines Agency, for ongoing or unpublished trials. The systematic literature search was conducted on February 11, 2024. No restrictions on language or document type were put in place. We have shown the detailed search strategies in the supplemental information (Table S2).

Inclusion criteria for this study included the following: (1) randomized controlled trials that compared outcomes between the biological therapies group and placebo group in EoE and other non-EoE EGIDs patients; (2) no age, sex, or regional limits; (3) published as complete articles with data that can be extracted; and (4) no language restrictions. Among the excluded were the following: (1) case reports/series, nonoriginal studies (letters, reviews, editorials); (2) animal or in vitro studies; (3) studies with no access to full-text content, pure abstracts of papers; (4) studies without available data can be extracted; (5) noncomparative studies; and (6) studies with duplicate and irrelevant data.

The literature review, data extraction, and quality evaluation were carried out separately by two reviewers. They used a standardized data extraction sheet to excavate pertinent data from each study that was qualified and then cross-checked the outcomes. A third reviewer was present for discussion if there were any discrepancies. This review was carried out following the Cochrane guidelines. The following information was systematically extracted from the data using a standardized data collection table: (1) generic information – comprising title, authors, country, publication year, and study design; (2) baseline characteristics – number of participants, age, sex, course of the disease, intervention facilities for the experimental and control groups, route of administration, duration of treatment/follow-up, and histologic and symptomatic response criteria; (3) outcomes – histologic, symptomatic and endoscopic improvement; and (4) adverse events.

Although we have integrated the main outcomes of the various studies, the definitions of histological improvement and symptom relief used in different trials are somewhat heterogeneous, for example, four RCTs defined histological response as <6 eosinophils/HPF, one as ≤6 eosinophils/HPF, one as <5 eosinophils/HPF, while another defined histological response as a 75% reduction in eosinophil counts in both proximal and distal esophageal biopsy specimens. Similarly, the scoring systems used for symptom relief differed, Hirano et al. [22] used the SDI score, Rothenberg et al. [23] used the MDQ questionnaire, and Dellon et al. [24] chose the EEsAI score, among others, which are specifically noted in Table 1.

We performed a rigorous examination for quality assessment and bias risk of the included RCTs. The evaluation content included the following aspects: selection bias, performance bias, detection bias, attrition bias, reporting bias, and other biases. All research findings were evaluated according to these standards using the terms ‘low risk’, ‘unclear risk’, and ‘high risk’.

Data analysis of the study was carried out by Review Manager 5.4 and STATA 17.0. Dichotomous outcomes were reported as risk ratios (RR) with 95% CI, while continuous outcomes were revealed using the standardized mean difference (MD, SMD) with 95% CI. We used I² values to test the statistical heterogeneity, and the fixed effect model was chosen if p ≥ 0.1 and I²≤50% (statistical heterogeneity was considered to be appropriate). Subgroup analysis and sensitivity analysis were adopted to investigate the source of heterogeneity if p < 0.1 and I²>50% (statistical heterogeneity exists between the results). When it was impossible to identify the cause of heterogeneity, the random effect model was used for data analysis. The study’s publication bias was depicted using Harbord’s modified test and modified Galbraith plots. Harbord’s modified test revealed a significant difference with p < 0.05.

If we encountered incomplete data in some studies, we endeavored to contact the original researchers to make up for the missing data wherever possible.

3. Results

A total of 3,237 records were retrieved in our initial search, and 784 documents were removed as duplicates, leaving 2,434 records. A total of 2,411 records were disqualified after titles or abstracts were examined. A total of 23 studies were identified; 16 were ultimately eliminated because they failed to match the inclusion criteria (Table S3). One of the studies (Dellon et, al. 2022) consisted of two parts, A and B, for the sake of data extraction and analysis, and we considered them as two different trials. The flowchart of the literature screening procedure is depicted in Figure 1. These studies were all reported in English. Our meta-analysis included randomized controlled trials published up to February 11, 2024 [22–28], involving 792 patients with an age range of 5 to 75 years old, all the patients had been diagnosed with EoE. Participants in these trials were administered different biologics, 513 of whom received the treatment (experimental group) and 280 of whom received a placebo (control group), and the treatment period ranged from 12 to 24 weeks. The baseline data of each trial considered are summarized in Table 1.

The methodological quality items of the included studies as determined by the Cochrane Risk of Bias Assessment Tool are shown in Figure 2. Because each trial was double-blinded, performance bias was assessed as low for all trials. Only one study was evaluated as having an uncertain risk for this item since it did not specify the specific method of randomized sequence generation. High risk was assigned to two studies for this item because they failed to indicate allocation concealment. In addition, two studies were given a rating of questionable risk of detection bias due to ambiguous result assessment blinding techniques. One study’s attrition bias risk was graded as uncertain because there was insufficient data on the baseline and outcome variables. Three studies were classified as having an unknown risk of reporting bias because they lacked precision in disease diagnostic criteria and effectiveness criteria.

4. Discussion

This study comprehensively reviewed published double-blind randomized controlled trials investigating the efficacy and safety of biologics in treating EoE and non-EoE EGIDs. The primary objective was to assess the therapeutic effects of biologics and offer insights into potential advancements in this area. However, despite extensive search efforts, only seven articles focusing on EoE met the inclusion criteria. These studies encompassed interventions involving five distinct biological agents. Notably, while two trials explored biologics for non-EoE EGIDs and demonstrated some therapeutic benefits, these interventions (benralizumab and lirentelimab) have been discontinued by the manufacturer due to insufficient efficacy in the Phase III trials [29,30]. Overall, our meta-analysis revealed that biologics effectively alleviated clinical symptoms in EoE patients (RR 1.20 [CI 1.02–1.41]; p = 0.03) and induced histological improvements (RR 4.32 [CI 3.14–5.93]; p < 0.00001) compared to placebo. Although differences are observed in symptomatic improvement among various biologic classes, our study provides compelling evidence supporting biologics as a promising therapeutic avenue for EoE.

Despite EoE and other non-EoE EGIDs being diagnosed more frequently over time as a consequence of escalating incidence rates and enhanced recognition, the heterogeneity of their symptoms and the complexity of pathogenesis still make treating them a difficult task. Most drugs currently lack specific indications or approvals for use in EoE and other non-EoE EGIDs [12,31]. Primarily recommended therapies for EoE encompass PPIs, corticosteroids, dietary adjustments, or surgical intervention, while oral corticosteroids, dietary modifications, and anti-allergy medications are likewise suggested for non-EoE EGIDs. Nonetheless, the substantiating evidence for these approaches is limited due to concerns regarding efficacy gaps and potential medication-related adverse effects, and data on dietary and oral medications for long-term treatment of EoE or other non-EoE EGIDs are lacking [32–34]. To address these demands and foster innovative treatment strategies, an increasing number of clinical trials initiated in recent years to evaluate biologics for EoE or other non-EoE EGIDs are actively ongoing. Biologics were not endorsed as a treatment for EoE in the guidelines, which were published in 2017 and 2020 [33,35]. It is noteworthy that the US Food and Drug Administration (FDA) has approved dupilumab for EoE treatment in 2022 [36]. In past meta-analyses, only a very limited number of articles have presented evaluations of the efficacy of biologics for EoE, and a small subset of biologics-related RCTs have been considered [37–39]. Therefore, a thorough meta-analysis that statistically summarizes all current data on the use of biologics to treat EoE is still absent.

Biologics targeting IL-5 have been exploited and investigated in EoE and other non-EoE EGIDs, such as mepolizumab, reslizumab, and benralizumab. Mepolizumab and reslizumab bind directly to the IL-5 molecule, interfering with its attachment to IL-5Rα expressed on the eosinophil cell surface. While Benralizumab targets the eosinophilic IL-5Rα chain expressed on the granulocyte membrane, blocking IL-5 access to its receptor [40,41]. However, Benralizumab has been discontinued by its manufacturer presently. In a recent randomized controlled trial (RCT) assessing Benralizumab, it was observed that although the medication alleviates hypereosinophilia by eliminating all circulating eosinophils in EoG, patients continued to exhibit persistent signs, symptoms, and biomarkers associated with EoG [30]. Our study revealed that EoE patients’ histologic remission and endoscopic improvement were both significantly enhanced by treatment with anti-IL-5 monoclonal antibodies. Nevertheless, the same effects were not observed in symptom evaluation. One possible explanation for the disparities in histological and symptomatic responses resulting from anti-IL-5 monoclonal antibody treatment is that the RCTs we included used different criteria to assess symptomatic responses. One example is the EoE Symptom Activity Index (EEsAI) score, which focuses on the symptoms of dysphagia, whereas the physician’s eosinophilic oesophagitis global assessment scores were used for analyses that focus on the overall assessment of patient symptoms. However, there is no consensus on which scores to use to assess these patients, so studies of the different scores used were summarized. Indeed, the assertion posited by Kliewer et al. in their study presents another plausible interpretation suggesting that the pathogenesis of EGIDs may not solely hinge on eosinophilic mechanisms. Instead, it underscores a predilection for a comprehensive type 2 immune targeting approach.

Emerging clinical and experimental evidence underscores that excess IL-4 and IL-13 are central drivers in EoE [42,43]. A monoclonal antibody called dupilumab targets IL-4R and functions by recognizing and blocking the common receptor of the IL-4 and IL-13 molecules [28]. It has been endorsed to treat several eosinophilic inflammatory diseases, including EoE [14]. Additionally, the study included two anti-IL-13 monoclonal antibodies, QAX576 and RPC4046. By directly interacting with the IL-13 molecule, QAX576 prevents both the synthesis of eotaxin and IL-13 activity. RPC4046 acts by inhibiting the binding of IL-13 to the IL-13Rα receptor subunits [44]. Our analysis indicates that biologics aiming to block IL-4/IL-13 have achieved statistically significant improvements in histological, clinical, and endoscopic aspects of treating EoE. The latest research suggests that dupilumab also exhibits higher long-term efficacy in EoE [45]. Currently, no published RCT are investigating the efficacy of biologics in treating non-eosinophilic EGIDs. Only one relevant clinical trial (NCT03678545) is underway, highlighting the scarcity of research in this area.

Siglec-8, an inhibitory receptor on human eosinophils and mast cells, has manifested as a prospective therapeutic target in allergy and inflammatory diseases. There are currently two kinds of Siglec-8-targeted biologics, AK001 and AK002, in the clinical development stage [40,46,47]. AK002 is a nonfucosylated IgG1 anti-Siglec-8 monoclonal antibody. Findings from a phase II clinical trial demonstrated that lirentelimab, an anti-Siglec-8 antibody, led to decreased gastrointestinal eosinophils and alleviated symptoms in individuals with EoG and eosinophilic duodenitis (EoD) [29]. However, data from the phase III clinical trial indicated that while patients with EoE, EoG, and EoD exhibited improved histology, symptomatic relief was not observed. It is noteworthy that lirentelimab was discontinued due to insufficient clinical efficacy. Several clinical studies on AK002 are still in progress, and further follow-up studies are needed in the future.

All of the biologics included in our study exhibited a generally well-tolerated profile, with a low incidence of adverse events during treatment, most of which were mild to moderate. Notably, adverse events during therapy occurred approximately equally frequently between the biologics and control groups. The most prevalent reported adverse events included nonsevere injection site reactions, nasopharyngitis, headache, and cough. During the double-blind period, only a negligible number of patients experienced severe side effects that were almost unrelated to the treatment. Furthermore, no fatalities were reported across all studies, and vital signs and other ancillary tests of patients showed no clinically meaningful alterations. Given the chronicity of EoE, the prolonged administration of both systemic and topical corticosteroids is imperative, with substantial side effects, and withdrawal may lead to the recurrence of symptoms [48]. In this context, biological agents have demonstrated pronounced advantages.

A primary limitation of this study is the limited scope of included trials, all of which exclusively focused on EoE. Consequently, there is a notable paucity of data concerning the efficacy of biologics in treating non-EoE EGIDs. We were unable to further categorize the treated group by age range due to the small number of trials. Moreover, the heterogeneity among the articles was exacerbated by the different types and doses of biologics. Although we performed subgroup analyses based on different types of biologics, the number of studies related to each biologic was too small to conduct further dose-response studies. In addition, there was heterogeneity in the diagnosis of EoE among the included studies. While in the majority of studies, EoE was defined as having eosinophil counts of at least 15 per high-power field (HPF), a few studies used higher thresholds, such as at least 20 or 24 eosinophils per HPF. This variation could introduce selection bias by excluding patients with milder cases from the study. Finally, the criteria for outcome indicators exhibited variability across the studies, including variations in symptom scores and definitions of histological improvement. The results may be affected by greater heterogeneity.

In conclusion, our study results showed that biological therapies for EoE have proven effective in reducing eosinophil counts, alleviating clinical symptoms, mitigating endoscopic abnormalities, and maintaining a favorable safety profile. Novel biological agents are considered convenient alternatives and merit additional research, particularly given the limited efficacy of both medical and dietary treatments for EoE and other non-EoE EGIDs. Nevertheless, given the constraints of this study, the determination of the optimal dose, duration, and long-term safety of biological agents lacks substantial support from more definitive data. More high-quality trials, preferably incorporating clinical assessments with validated and standardized scoring systems, should be conducted in the future to investigate the ideal dosage of biologics as well as identify the appropriate patient population for treatment.

Supplementary Material

Funding Statement

This work was supported by the Key Research and Development Projects of Shaanxi Province (S2023-YF-YBSF-1670).

Author contributions

Beibei Zeng and Doudou Jia conceptualized and designed the study, conducted the literature search, analyzed data, drafted the manuscript, and reviewed and revised the manuscript; Shengen Li and Xuna Liu reviewed literature, extracted data, evaluated quality, and revised the manuscript; Boxu Zhu, Yanqi Zhang, and Yan Zhuang designed the data collection instruments, analyzed data, and revised the manuscript; Dr Fei Dai conceptualized and designed the study, coordinated and supervised data collection, and critically reviewed the manuscript for important intellectual content. The final manuscript as submitted was approved by all authors and they agreed to accept responsibility for all aspects of this work.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Trial registry

PROSPERO; No. CRD42023440744; URL: http://www.crd.york.ac.uk/PROSPERO.

Data availability statement

The corresponding author is available with the study’s data upon reasonable request.