Pharmacological Interventions for Eosinophilic Esophagitis
Pharmacological Interventions for Eosinophilic Esophagitis
The search yielded 422 unique references of which 405 did not meet the inclusion criteria, resulting in 17 RCTs for the systematic review (Figure 1, Table S1 ). Thirteen originated in North America and four in Switzerland. Eleven studies were conducted in adults, five in children (age <18) and one in both children and adults. Within the 17 included studies, the female/male ratio was ~1:3. We also had more children (483 children, 381 adults) with a mean age for adults of 35.4.
(Enlarge Image)
Figure 1.
Flow chart of the systematic review and meta-analysis.
Although all the studies included patients with oesophageal symptoms, inclusion criteria differed between studies regarding the histopathological diagnosis cut-off of peak eosinophil count and the number of oesophageal biopsies for diagnosis.
Among the studies which reported symptom outcomes (n = 11), the evaluation of oesophageal symptomatology was assess by different scales. Studies conducted in adults used scores that are mainly based on dysphagia severity. Four studies used the Mayo Dysphagia Questionnaire (MDQ) and two studies used Physician's Global Assessments. The rest of the studies assessed symptoms with individualised scales. In contrast to adults, children had a wider spectrum of symptoms while dysphagia was less common as compared to adults. Children who did not have dysphagia were included in these studies, however, this made the dysphagia scale inappropriate for the paediatric population. Gupta et al. and Dohil et al. developed a clinical symptom score for children. The score is based on the physician assessment of six cardinal symptoms using a questionnaire filled by the subject or the caregiver.
According to the updated ACG guidelines, the histological diagnosis should be made by getting 2–4 biopsies from both the proximal and distal oesophageal with results showing a peak eosinophil equal or greater than 15 eos/hpf, in addition to obtaining biopsies from the antrum and duodenum. All studies reported peak count of 15 eos/hpf or more in the oesophageal biopsy. The histological cut-off for EoE diagnosis ranged between ≥15 and >24 eos/hpf and mean peak ranged between 35 and 126 eos/hpf. Only nine studies, however, met the ACG criteria in term of getting biopsies from both the proximal and distal oesophageal. Only four of the 17 studies reported acquisition of biopsies from the antrum and duodenum to exclude other causes of eosinophilia.
Histological remission varied between decrease in eosinophil count, <0–6 eos/hpf or 90% decrease in eosinophils compared to baseline. Seven studies considered decrease in the peak eosinophil count to 5 or less post-treatment as a complete response. The ACG defined histological remission as eosinophilic inflammation improvement.
A trial of PPI treatment prior to inclusion was mandated in six studies (two topical steroids, four biological drugs). Six studies allowed patients to be on PPI before randomisation. The rest of the studies excluded patients who were treated with PPI or did not have clear criteria. Two studies excluded secondary causes of oesophageal eosinophilia, whereas the rest did not report whether secondary causes of oesophageal eosinophilia were excluded or not.
Twelve studies evaluated topical steroids as a treatment. Topical steroids were either swallowed budesonide or fluticasone in a nebuliser form or oral viscous form. The dose for fluticasone ranged between 110–880 μg twice per day and budesonide dose ranged between 0.5–2 mg/day (Table S1 ). The control groups differed among studies (Table S1 ). Eight studies had placebo, two studies used PPI, and oral prednisone was used in one study. Schaefer et al. randomised patients to receive either systemic oral prednisone or topical swallowed fluticasone metered-dose inhaler (MDI), whereas Dellon et al. compared oral viscous budesonide (OVB) to the nebulizer form budesonide. Straumann et al. evaluated long-term budesonide on maintaining remission compared to placebo.
Five studies evaluated biological drugs. Spergel et al.and Straumann et al. randomised patients to monoclonal antibody anti-interleukin-5 (anti-IL-5) vs. placebo. The anti-IL-5 drugs used in the trials were Mepolizumab (0.55–10 mg/kg) and Reslizumab (1–3 mg/kg). Both Clayton et al. and Fang et al. randomised patients into anti-IgE (Omalizumab) and placebo. Another study by Straumann et al. 2012 randomised patients to anti Th2 cells (OC000459) vs. placebo. OC000459 is a selective CRTH2 antagonist, which blocks the recruitment and activation of Th2 cells and eosinophils.
Treatment duration ranged from 2 to 12 weeks with topical steroids and between 2 and 16 weeks with anti-IL-5. Straumann et al. compared treatment with topical Budesonide to placebo for 50 weeks as a maintenance therapy after achieving remission.
American College of Gastroenterology defined symptomatic response as resolution of EoE symptoms. Although symptomatic response was different between the studies, 11 of 17 medical intervention provided data regarding symptomatic response. The ACG defined histological improvement as eosinophilic inflammation improvement. Each study had a different definition and cut-off for histological remission, but all of them conformed with the ACG definition (improvement in eosinophilic inflammation; Table S1 ). Nine studies showed significant histological remission and only four studies showed significant symptomatic response with topical steroids. In the one study, Schaefer et al. compared oral prednisone with topical fluticasone, there was no difference between oral prednisone and topical steroid with regard to symptomatic, histological remission and relapse risk. However, topical fluticasone had fewer systematic side effects (40% in the prednisone compared to 0% in the fluticasone). Oesophageal candidal overgrowth was seen in 15% patients in the fluticasone arm and none in the prednisone arm. Regarding formulations, oral viscous budesonide was more effective than nebulised in reducing the number of eosinophils, but both forms were similar and effective in terms of inducing symptomatic improvement and in side-effect rates.
In the five studies that evaluated biological agents, complete histological remission as defined by decrease in eosinophil count to five or less eos/hpf or significant reduction in eosinophil count was not achieved. However, there was a significant improvement in eosinophilic inflammation with the use of biological drugs. None of these studies was able to show an improvement in achieving symptomatic remission. The dose of the biological drug did not affect remission rate.
Of the 17 RCTs, four studies had a high risk for bias, four had unclear risk for bias and nine were low risk (Figure S1 ). The major problem with the high risk of bias studies was blinding.
Results
Study Characteristics
The search yielded 422 unique references of which 405 did not meet the inclusion criteria, resulting in 17 RCTs for the systematic review (Figure 1, Table S1 ). Thirteen originated in North America and four in Switzerland. Eleven studies were conducted in adults, five in children (age <18) and one in both children and adults. Within the 17 included studies, the female/male ratio was ~1:3. We also had more children (483 children, 381 adults) with a mean age for adults of 35.4.
(Enlarge Image)
Figure 1.
Flow chart of the systematic review and meta-analysis.
Although all the studies included patients with oesophageal symptoms, inclusion criteria differed between studies regarding the histopathological diagnosis cut-off of peak eosinophil count and the number of oesophageal biopsies for diagnosis.
Symptom Outcomes Definitions
Among the studies which reported symptom outcomes (n = 11), the evaluation of oesophageal symptomatology was assess by different scales. Studies conducted in adults used scores that are mainly based on dysphagia severity. Four studies used the Mayo Dysphagia Questionnaire (MDQ) and two studies used Physician's Global Assessments. The rest of the studies assessed symptoms with individualised scales. In contrast to adults, children had a wider spectrum of symptoms while dysphagia was less common as compared to adults. Children who did not have dysphagia were included in these studies, however, this made the dysphagia scale inappropriate for the paediatric population. Gupta et al. and Dohil et al. developed a clinical symptom score for children. The score is based on the physician assessment of six cardinal symptoms using a questionnaire filled by the subject or the caregiver.
Histological Outcomes Definitions
According to the updated ACG guidelines, the histological diagnosis should be made by getting 2–4 biopsies from both the proximal and distal oesophageal with results showing a peak eosinophil equal or greater than 15 eos/hpf, in addition to obtaining biopsies from the antrum and duodenum. All studies reported peak count of 15 eos/hpf or more in the oesophageal biopsy. The histological cut-off for EoE diagnosis ranged between ≥15 and >24 eos/hpf and mean peak ranged between 35 and 126 eos/hpf. Only nine studies, however, met the ACG criteria in term of getting biopsies from both the proximal and distal oesophageal. Only four of the 17 studies reported acquisition of biopsies from the antrum and duodenum to exclude other causes of eosinophilia.
Histological remission varied between decrease in eosinophil count, <0–6 eos/hpf or 90% decrease in eosinophils compared to baseline. Seven studies considered decrease in the peak eosinophil count to 5 or less post-treatment as a complete response. The ACG defined histological remission as eosinophilic inflammation improvement.
Exclusion of Other Eosinophilic Enteropathies
A trial of PPI treatment prior to inclusion was mandated in six studies (two topical steroids, four biological drugs). Six studies allowed patients to be on PPI before randomisation. The rest of the studies excluded patients who were treated with PPI or did not have clear criteria. Two studies excluded secondary causes of oesophageal eosinophilia, whereas the rest did not report whether secondary causes of oesophageal eosinophilia were excluded or not.
Medical Therapies Evaluated
Twelve studies evaluated topical steroids as a treatment. Topical steroids were either swallowed budesonide or fluticasone in a nebuliser form or oral viscous form. The dose for fluticasone ranged between 110–880 μg twice per day and budesonide dose ranged between 0.5–2 mg/day (Table S1 ). The control groups differed among studies (Table S1 ). Eight studies had placebo, two studies used PPI, and oral prednisone was used in one study. Schaefer et al. randomised patients to receive either systemic oral prednisone or topical swallowed fluticasone metered-dose inhaler (MDI), whereas Dellon et al. compared oral viscous budesonide (OVB) to the nebulizer form budesonide. Straumann et al. evaluated long-term budesonide on maintaining remission compared to placebo.
Five studies evaluated biological drugs. Spergel et al.and Straumann et al. randomised patients to monoclonal antibody anti-interleukin-5 (anti-IL-5) vs. placebo. The anti-IL-5 drugs used in the trials were Mepolizumab (0.55–10 mg/kg) and Reslizumab (1–3 mg/kg). Both Clayton et al. and Fang et al. randomised patients into anti-IgE (Omalizumab) and placebo. Another study by Straumann et al. 2012 randomised patients to anti Th2 cells (OC000459) vs. placebo. OC000459 is a selective CRTH2 antagonist, which blocks the recruitment and activation of Th2 cells and eosinophils.
Treatment duration ranged from 2 to 12 weeks with topical steroids and between 2 and 16 weeks with anti-IL-5. Straumann et al. compared treatment with topical Budesonide to placebo for 50 weeks as a maintenance therapy after achieving remission.
Outcome Improvements: Qualitative Summary of the Evidence
American College of Gastroenterology defined symptomatic response as resolution of EoE symptoms. Although symptomatic response was different between the studies, 11 of 17 medical intervention provided data regarding symptomatic response. The ACG defined histological improvement as eosinophilic inflammation improvement. Each study had a different definition and cut-off for histological remission, but all of them conformed with the ACG definition (improvement in eosinophilic inflammation; Table S1 ). Nine studies showed significant histological remission and only four studies showed significant symptomatic response with topical steroids. In the one study, Schaefer et al. compared oral prednisone with topical fluticasone, there was no difference between oral prednisone and topical steroid with regard to symptomatic, histological remission and relapse risk. However, topical fluticasone had fewer systematic side effects (40% in the prednisone compared to 0% in the fluticasone). Oesophageal candidal overgrowth was seen in 15% patients in the fluticasone arm and none in the prednisone arm. Regarding formulations, oral viscous budesonide was more effective than nebulised in reducing the number of eosinophils, but both forms were similar and effective in terms of inducing symptomatic improvement and in side-effect rates.
In the five studies that evaluated biological agents, complete histological remission as defined by decrease in eosinophil count to five or less eos/hpf or significant reduction in eosinophil count was not achieved. However, there was a significant improvement in eosinophilic inflammation with the use of biological drugs. None of these studies was able to show an improvement in achieving symptomatic remission. The dose of the biological drug did not affect remission rate.
Of the 17 RCTs, four studies had a high risk for bias, four had unclear risk for bias and nine were low risk (Figure S1 ). The major problem with the high risk of bias studies was blinding.