Trial design

Patients were included in a 6-month randomised double-blind placebo-controlled study and allocated to treatment with FMT capsules or placebo capsules. Twenty-five capsules were consumed while fasting every morning for 12 days. Before the first treatment, the participants had a bowel cleansing with Picoprep performed corresponding to the procedure before a colonoscopy.

All patients were seen for study visits at baseline, 1, 3 and 6 months, where they completed the IBS-severity scoring system (IBS-SSS)15 and IBS-specific quality of life (IBS-QoL).16 17 Additionally, patients were asked to keep a daily diary including Bristol Stool Form Scale,18 symptoms, use of laxatives and side effects of the treatment, if any.

The primary end point was to evaluate the reduction of IBS-SSS in the treatment group compared with the placebo group at 3 months. Secondarily, we evaluated change in IBS-QoL scores at 3 months and changes in microbiota diversity before and after FMT.

Participants

Adult patients with IBS (aged 18–60 years) were recruited between October 2016 and December 2016 from the Department of Gastroenterology, Aleris-Hamlet Hospitals Copenhagen, Søborg, Denmark and Department of Gastroenterology, Copenhagen University Hospital Hvidovre, Copenhagen, Denmark. Patients were diagnosed with IBS according to Rome III criteria.2

In addition, criteria were as follows:

Inclusion criteria:

• Moderate-to-severe disease activity (IBS-SSS ≥175);

• Able to read and speak Danish;

• Normal colonoscopy (performed within 1 year) if the patient was ≥40 years or if the patient had blood in stool.

Exclusion criteria:

• Other chronic GI disease;

• Faecal sample positive for enteropathogenic microorganisms;

• Positive screening for HIV, HBV or HCV antibody;

• Surgical interventions in the GI region (except for appendectomy, hernia repair, cholecystectomy and gynaecological and urological procedures);

• Psychiatric disorder;

• Faecal calprotectin ≥50 mg/kg;

• Abuse of alcohol or drugs;

• Medications other than birth control pills, hormone supplements, allergies/asthma agents, blood pressure and cholesterol-lowering agents, proton pump inhibitors and non-prescription medicines;

• Abnormal screening biochemistry;

• Abnormal colonoscopy findings;

• Pregnant, planned pregnancy or breastfeeding females;

• Ingestion of probiotics or antibiotics <8 weeks before the inclusion.

Furthermore, all patients were subclassified in three different IBS subtypes: constipation-predominant (IBS-C), diarrhoea-predominant (IBS-D) or alternating periods of constipation and diarrhoea (IBS-M).18 Demographic information was obtained in all participants and they reported their current use of medications and completed questionnaires to characterise their symptom severity and bowel habits. The patients were asked to live as they used to through the half year study period.

Only a few patients fulfilling inclusion criteria declined participation.

Donors

Four faecal donors were recruited to this study. Once recruited, the donors were instructed to keep up a healthy lifestyle during the collecting period. They were all screened according to guidelines19 20 and were recruited according to the following criteria:

Inclusion criteria:

• Aged between 18 and 45 years;

• Previously and currently healthy;

• Normal weight (body mass index (BMI) between 18.5 and 24.9 kg/m²);

• Normal bowel movements (defined as 1–2 per day and type 3–4 at Bristol Stool Form Scale);

• No medication consumption.

Exclusion criteria:

• Known or high risk of infectious diseases such as HIV, HAV, HBV or HCV;

• Positive stool sample for C. difficile toxin, parasites or other enteropathogens;

• Antibiotic treatment in the past 6 months;

• Abuse of alcohol or drugs;

• Smoking;

• Tattoo or body piercing within the last 6 months;

• Allergy, asthma or eczema;

• Family history of GI diseases, cancer, diabetes, obesity, autoimmune diseases, allergy, asthma, eczema, cardiovascular diseases, neurologic or mental illnesses;

• Participation in high-risk sexual behaviours;

• Born by caesarean section.

Sample preparation

Donors were equipped with 500 mL bottles of oxygen reduced sterile saline (0.9% NaCl), which they were instructed to keep refrigerated. Immediately after producing the sample, donors were instructed to cover the sample in the oxygen reduced saline to protect the sample from oxygen and to deliver it to our facility within 1 hour. Thereafter, the sample was stored at 5°C and processed in the laboratory no more than 3 hours later from the time of delivery. Samples with saline were homogenised manually, using a 400 mL BagPage XR from Interscience with a 250 µm filter to remove undigested fibrous material and centrifuged at 3000x g for 20 min at room temperature to remove the added saline. In this way we obtained a more concentrated product. The supernatant was discarded and the pellet was mixed with glycerol as a cryoprotectant to a final concentration of 30% glycerol. The effect of glycerol 30% in the FMT capsules was for freeze protection. With a lower amount of glycerol our capsules dissolved at −20°C. Most of the process was done in an anaerobic environment, using Argon gas to protect the sample. Faecal matter was briefly exposed to oxygen only while being transferred between containers. Finally, the samples were frozen at −20°C. Once donors had passed the second screening, all faecal samples were mixed into one batch for the entire experiment before being double encapsulated using Capsugel DR Caps size 0 and 00. One daily dosage of 25 such capsules contains approximately 12 g of the frozen faecal matter, which was derived from approximately 50 g of fresh faeces.

Interventions

FMT were in capsule form, packaged in plastic bottles. FMT and placebo capsules were identical in appearance: form, colour and size. Placebo capsules where made from saline, glycerol and food colouring E150. Also the placebo contained 30% glycerol. The identity of the capsules was unknown to participants, researchers and primary investigators. Participants were instructed to consume orally 25 capsules per day in the morning with water. It was allowed to eat an hour after capsule intake to ensure that most of the capsules had passed the stomach. The first daily dose was taken under supervision at the Department of Gastroenterology, Aleris-Hamlet Hospitals Copenhagen, where the study visits took place.

Faecal sample collection

Faecal samples were longitudinally collected from patients at baseline before bowel cleansing, 3 days after FMT treatment had stopped, and 1, 3 and 6 months after inclusion. The collected fresh faeces were stored in RNA later by the patients and brought to the hospital at study visits.

Library preparation and sequencing

Total DNA was extracted from 0.25 g of faecal samples using a PowerLyzer PowerSoil DNA Isolation Kit (MoBio 12 855–50) according to the manufacturer’s instructions. The V3-V4 hypervariable regions of 16S rRNA genes were amplified using primers 341F and 806R and indexed Illumina compatible 16S amplicon libraries were prepared as described elsewhere.21 Paired end sequencing (2×301 bp) was performed on the Illumina MiSeq platform with the MiSeq Reagent kit V.3.

Sequence processing and microbiomics analysis

Sequencing reads were demultiplexed using bcl2fastq V.2.17.1.14 (Illumina). Primer sequences were striped from the 5’ ends of each read in a pair and reads without discernible primer sequences were discarded using a custom Biopython script. Reads were processed with the UPARSE pipeline using usearch V.10.0.240_i86linux32 to generate an operational taxonomic unit (OTU) table at 97% granularity.22 The UPARSE-pipeline was modified, replacing the ‘read quality filtering’ and ‘length trimming’ steps with an alternative quality filtering (usearch -fastq_filter -maxee 1.0).23

Usearch was also used to construct an OTU tree and assign taxonomy to the OTUs’ centroids24 25 using the ribosomal database project reference 16S training set with species names (V.16) as a reference database. Taxonomy assignments were added to the OTU table with a custom Python script and the OTU table with taxonomy information was converted into biom format using biom.26 Diversity metrics were calculated using the Qiime V.1.9.1 core_diversity_analyses.py27 pipeline using a sampling depth of 5000 and default parameters otherwise.

OTU abundances in pairs of samples were assessed for significant differences using the Mann-Whitney U test, corrected for multiple sampling with the Benjamini-Hochberg method (FDR=5%). Correlations between IBS-SSS and OTU abundances and correlations between IBS-SSS and α-diversity were assessed using Spearman-Rank correlation analysis, corrected for multiple sampling with the Benjamini-Hochberg method (FDR=5%). Significant differences in α-diversity (chao1 metric) between treatment groups were assessed using the Mann-Whitney U test (p≤0.05). Differences in β-diversity distances were assessed using a Mann-Whitney U test (p≤0.01). SourceTracker was used to infer the proportions of microbial communities that come from possible source environments.28

Ethics

The study was performed in accordance with the requirements of Good Clinical Practice and the Revised Declaration of Helsinki. The study was registered in www.clinicaltrials.gov (NCT02788071). All participants provided written informed consent to participate after verbal and written information about the study. Participants could discontinue at any time point on their request. As FMT, according to the Danish Health Authority, is not considered as a pharmaceutical, no authorisation by the Danish Medicines Agency was required. All authors had access to the study data and reviewed and approved the final manuscript.

Outcomes

Statistical methods

All statistical analyses were done in RStudio (RStudio Team (2016). RStudio: Integrated Development for R. RStudio, Boston, Massachusetts, USA; http://www.rstudio.com/). The number, percentage, mean, SD, mean difference and 95% CIs were reported. The IBS-QoL score was transformed into a 0–100 scale using the formula: total score=(sum of the items−34/170)×100. Student’s t-test was used to determine the difference between placebo and FMT. A paired t-test and a repeated measure analysis of variance (rm-ANOVA), was used to determine, within each group, the difference between inclusion and 1, 3 and 6 months follow-up. All tests were done for both IBS-SSS and IBS-QoL. Spearman’s test was used to determine the correlation between IBS-SSS and IBS-QoL. A score between 0 and 0.19 was interpreted as very weak, 0.20 and 0.39 as weak, 0.40 and 0.59 as moderate, 0.60 and 0.79 as strong and 0.8 and 1 as very strong.

Two-way ANOVA was used to determine factors associated with decrease in either IBS-SSS or IBS-QoL score. If interaction was found between variates, a linear regression model was used instead, controlled by quantile-quantile plot of residuals.

Furthermore, a cut-off of 50 points15 reduction after 3 months in the IBS-SSS was used to distinguish between ‘effect’ and ‘no effect’. This end point was analysed using the χ² test. A logistic regression was used to determine associated factors for ‘effect’. Age, gender, BMI, previous or concurrently used IBS medical therapy, birth by caesarean section, breast feeding, IBS subtype, previous attempt with change in diet and weight loss were used as independent variables.

Measures

IBS disease severity was measured by using the IBS-SSS questionnaire,15 which includes five items on a 0–100 mm visual analogue scale with total score ranging from 0 to 500 mm. Question 1: severity of abdominal pain, question 2: frequency of abdominal pain, question 3: severity of abdominal distension, question 4: dissatisfaction with bowel habits and question 5: interference with quality of life.

Alterations in quality of life during the study were measured by IBS-QoL questionnaire, which consists of 34 items, each with a 5-point response scale. The 34 items are based on the following eight variables: health worries, food avoidance, body image, dysphoria, interference with activity, social reactions, sexual activity and relationships.17