Randomised Clinical Trial

Adjunctive Induction Therapy With Oral Effervescent Budesonide in Newly Diagnosed Coeliac Disease

Evan D. Newnham; Daniel Clayton-Chubb; Meena Nagarethinam; Patrick Hosking; Peter R. Gibson


Aliment Pharmacol Ther. 2021;54(4):419-428. 

In This Article

Materials and Methods


Participants for this study were recruited from local specialised coeliac disease clinics, dietitian practices and advertisements in coeliac-specific publications and newsletters between 2011 and 2014. Those aged between 16 and 75 years with newly diagnosed coeliac disease in whom duodenal histology showed a maximum severity of at least a Marsh 3A lesion (as demonstrated by at least four adequate biopsies in the second and/or third parts of the duodenum with or without a biopsy of the first part of the duodenum) and in whom the gastroscopy had been performed no more than 6 weeks prior to baseline were included. A negative pregnancy test was required for female participants. Exclusion criteria included treatment with a gluten-free diet for greater than 4 weeks; significant co-morbidity, including diabetes, untreated thyroid disease, heart failure or inflammatory bowel disease; abnormal liver function test as shown by alanine aminotransferase, aspartate aminotransferase or alkaline phosphatase greater than three times the laboratory reference upper limit of normal; the pre-morbid use of inhaled or systemic corticosteroids.


This was a randomised, double-blind, placebo-controlled, single-centre, parallel-group trial approved by the Eastern Health Research and Ethics Committee and was registered with the Australian and New Zealand Clinical Trials Registry (ACTRN12609000738224). Participants were randomised to treatment with 3 mg budesonide (delivered as an oral effervescent tablet dissolved in 250 ml water) or placebo similar in appearance and taste. A 1:1 ratio was realised by an independent statistician using randomly permutated blocks, resulting in a computer-generated randomisation list presented as sealed opaque envelopes. Recruitment and assigning of participants to the intervention were performed by EDN, who remained blinded to the intervention. Each participant in the budesonide arm received a total dose of 9 mg/day for 8 weeks given in divided doses of 3 mg each, followed by 6 mg/day for 1 week and 3 mg/day for 1 week. All participants were to undergo gastroscopy at weeks 8 and 52. At least four biopsies were obtained from the second part of the duodenum and at least two biopsies from the first part. A single, highly experienced pathologist (PH), blinded to both study visit and participant identity, undertook all assessments. These included the recording of Marsh score and a villous-height : crypt-depth ratio. To do this, measurements were attempted on every biopsy taken. The best orientated tissue fragment in each specimen that would give the most accurate measurement of villous length and crypt depth was selected, and then they were measured manually with a graticule in μm. If there were two to three well-orientated villi in a row, they were averaged down to one value in the final spread sheet. Poorly orientated fragments that were en face or did not show villi were ignored. At weeks 0, 4, 8, 10, 26, 38 and 52, participants completed multiple symptom assessments that included a 100-mm visual analogue scale to assess the severity of overall gastrointestinal symptoms and a single question regarding satisfaction with their level of symptoms. Fatigue was assessed by the Fatigue Impact Scale[17] and levels of anxiety and depression by the Hospital Anxiety and Depression Score.[18]

Safety Assessment

At each visit, the patients were directly interrogated for adverse events, and bloods were taken for routine biochemistry and haematology as well as fasting concentrations of serum cortisol blood glucose. At baseline and at 52 weeks, bone mineral density was assessed by dual-energy X-ray absorptiometry. Adverse events were divided into treatment-emergent events that occurred during the drug intervention period and post-treatment emergent events that occurred during the observation period from week 12.

Compliance Assessments

Adherence to the gluten-free diet was assessed utilising prospective food intake diaries, completed over the 7 days prior to the relevant study visit. The diary was reviewed and cross-checked by an experienced dietitian with advice provided to the participant. Adherence was rated according to the frequency of episodes of gluten intake, being 'excellent' if there were no episodes, 'good' if occurring less than two times over the first 3 months or eight times in 12 months, 'fair' if occurring up to once per week, and 'non-compliant' more than once per week. Compliance to the study medication was assessed via a tablet count of the returned tablets at each visit for the first 10 weeks of the study, and the percentage of tablets taken calculated.


The primary outcome was the proportion of patients achieving mucosal response (Marsh 1 or Marsh 0) at 8 weeks. Secondary outcomes included proportion of patients with mucosal remission (Marsh 0) by 8 and 52 weeks; proportion of patients with mucosal response at 52 weeks; degree of change in villous-height : crypt-depth ratio at weeks 8 and 52; degree of symptom change over the first 8 weeks; change in fatigue and measures of anxiety and depression at 8 weeks and 6 and 12 months; and change in bone density at 52 weeks.

Statistical Evaluation

As there were no previous data on early healing published, power calculations were not performed. However, it was planned to have 20 patients in each group. Dichotomous data were compared among treatment groups using Fisher's exact test. Continuous variables were compared using unpaired t test or Wilcoxon rank sum test, depending upon normality of data distribution. Safety variables were analysed descriptively only. For histological end-points, only patients undergoing endoscopy were analysed. For end-point and safety evaluation, the intention-to-treat (ITT) cohort was analysed, and missing values were imputed by the last-observation-carried-forward method. A post hoc analysis of baseline characteristics was performed in patients receiving 8-week histopathology to determine association with mucosal remission comparing those achieving mucosal remission at 8 weeks to those who did not. All statistical tests were performed two-sided with a 0.05 significance level. Where appropriate, a Bonferroni correction for multiple comparisons was performed.