https://orcid.org/0000-0003-1026-3173
Abstract
While surgery plays a pivotal role in the management of ileal Crohn’s disease, the risk of endoscopic recurrence following an ileocaecal resection can be greater than 65% within 12 months of surgery. More than 90% of patients with Crohn’s disease have a concomitant diagnosis of bile acid diarrhea following an ileal resection. This pilot study aimed to assess whether the use of bile acid sequestrants in patients with Crohn’s disease who have undergone a primary terminal ileal resection with concomitant bile acid diarrhea can alter the microbiome and prevent disease recurrence.
Patients with Crohn’s disease who underwent a primary terminal ileal resection and had symptoms of diarrhea within 1-3 months of surgery underwent 75SeHCAT testing for bile acid diarrhea. If positive (75SeHCAT ≤ 15%), patients were treated with colesevelam and stool samples were collected at 4 weeks, 8 weeks, and 6-12 months posttreatment. If negative (75SeHCAT > 15%), treatment was not given and were reviewed in the clinic as per local guidelines. All patients underwent a 6-12 month postoperative colonoscopy where further stool samples and mucosal biopsies were taken. Disease activity was established using the endoscopic Rutgeert’s score, with disease remission defined as Rutgeert’s score <i2 and disease recurrence ≥i2. 16S ribosomal RNA gene analysis was undertaken for the collected fecal and mucosal samples to assess α/β-diversity and microbial composition.
A total of 14 patients who completed the study, 10 of whom had a 75SeHCAT positive diagnosis of bile acid diarrhea and were started on treatment with colesevelam. Four patients did not require treatment as 3 were asymptomatic and 1 had a negative 75SeHCAT scan. Three of the fourteen patients had disease recurrence at their 6-12 month postoperative colonoscopy assessment, of which 1 patient was taking colesevelam and 2 patients were not taking colesevelam. A total of 44 fecal samples and 44 mucosal biopsies underwent 16S ribosomal RNA gene analysis to assess α/β-diversity and microbial composition. In the colesevelam treated patients there was no significant difference in α/β-diversity pre- and posttreatment. Pretreatment, the 3 most abundant bacterial classes in all patients were Bacteroidia, Clostridia, and Gammaproteobacteria. Following 6-12 months of treatment, out of the 9 patients on colesevelam, 5/9 (55.6%) had a reduction in Bacteroidia, 9/9 (100%) had an increase in Clostridia, and 7/9 (77.8%) had a reduction in Gammaproteobacteria. Of the 2 patients not given colesevelam, one showed a reduction in Bacteroidia, increase in Clostridia and a reduction in Gammaproteobacteria.
This small pilot study demonstrated that patients who were given colesevelam, were more likely to be in disease remission at their 6-12 months colonoscopy review compared with those not treated. Furthermore, treatment with colesevelam may have a role in altering the microbiome to help maintain remission states in postoperative Crohn’s disease. Larger mechanistic studies are now needed to confirm these findings and demonstrate statistical significance as well as investigate whether this benefit may be present even in those patients with 75SeHCAT negative disease.
Lay Summary
This pilot study demonstrated that patients with Crohn’s disease who underwent a primary terminal ileal resection and were given colesevelam were more likely to be in disease remission at their 6-12 month postoperative colonoscopy review compared with those not treated; there was a notable change in abundance in certain bacteria following treatment compared to their pretreatment microbiome.
- Disease recurrence is common following surgical intervention in Crohn’s disease, with a 70%-90% risk of patients developing endoscopic recurrence at their 1-year colonoscopy review, increasing to 80%-100% within 3 years.
- The gut microbiota is implicated in Crohn’s disease pathogenesis, with Clostridia associated with disease remission and Gammaproteobacteria associated with disease relapse.
- Changes in the microbial composition can result in alterations in the levels of gut microbiota-derived metabolites such as bile acids, which have also been implicated in Crohn’s pathogenesis.
- Patients on the bile acid sequestrant, colesevelam, were more likely to be in clinical and endoscopic remission 6 months following their primary terminal ileal resection.
- Patients treated with a bile acid sequestrant following their primary terminal ileal resection demonstrated an increased relative frequency of Clostridia and a reduction in Gammaproteobacteria.
- The use of colesevelam may help to alter the pathogenesis of postoperative Crohn’s disease by altering the microbiome and reducing the risk of relapse.
- Colesevelam has a good safety profile and does not carry the same side effects and long-term complications as the advanced therapies currently being offered to patients.
Introduction
Crohn’s disease (CD) is a complex chronic relapsing-remitting inflammatory bowel condition which has a preponderance towards the terminal ileum and large bowel.1,2 The treatment for CD has evolved over the years from symptom management to clinical and endoscopic remission.3,4 Surgery, however, continues to play a pivotal role in achieving disease control with at least one-third of patients requiring major abdominal surgery within 5 years of diagnosis and two-thirds requiring operative management at least once during the course of their disease.5-9 The LIR!C trial demonstrated that laparoscopic resection in limited terminal ileal disease (<40 cm) was a suitable alternative to conventional therapy and should be offered to selected patients as a first-line treatment.10 Unfortunately, disease recurrence is common following surgical intervention in CD, with a 70%-90% risk of patients developing endoscopic recurrence at their 1-year colonoscopy review, increasing to 80%-100% within 3 years.9,11-13 At least 35% of patients will undergo a second operation within 10 years of their initial surgery.14,15
The causes of CD are multifactorial and not fully understood but thought to arise from the interactions between our immune system, genetics, environment, and the microbiome.16 Research into the gastrointestinal microbiota has become increasingly recognized to play an important role in the development of CD with broad patterns emerging within the microbiota to help distinguish CD from other conditions (ie, ulcerative colitis).17 Changes in the microbial composition can result in alterations in the levels of gut microbiota-derived metabolites such as bile acids, which have also been implicated in the pathogenesis of CD.18 Bile acids can display both pro- and anti-inflammatory properties and its effects on the gut microbiota is influenced by several factors including bile acid concentration and exposure time, the receptors involved, the site of inflammation (ileum or colon) and bile acid transport and absorption processes.19 An excess of bile acids can result in bile acid diarrhea, which can affect up to 90% of patients with CD who have had their terminal ileum resected.20
Recent studies have demonstrated the role of bile acid sequestrants in CD in the induction of remission, improvement in symptoms, and overall quality of life.21,22 While cholestyramine and colestipol are the first-line treatment options for bile acid diarrhea, poor palatability due to the texture and taste of the resin powder often results in poor treatment compliance.23,24 Colesevelam, an unlicensed bile acid sequestrant, is available in tablet form and is generally better tolerated.25,26 Colesevelam differs structurally from conventional bile acid sequestrants due to its numerous hydrophobic side chains specifically added to enhance bile acid-binding.27 It thus forms nonabsorbable complexes with bile acids in the gastrointestinal tract, is not absorbed systemically, and is excreted unchanged from the gastrointestinal tract.28 Recently, the association between colesevelam treatment and changes in gut microbiome in patients with bile acid diarrhea has been suggested.29 Currently, however, there is no available literature that has explored the effect of colesevelam on bile acid receptors and subsequent microbiome modulation in patients with CD with concomitant bile acid diarrhea.
Thus, the primary aim of this study was to determine whether colesevelam can reduce the risk of disease recurrence in postoperative patients with CD. The secondary aim was to determine whether colesevelam can modify the microbiome.
Materials and Methods
Ethical Approval and Good Clinical Practice
The study was performed in accordance with the recommendations guiding physicians in biomedical research involving human subjects, adopted by the 18th World Medical Assembly, Helsinki, Finland 1964, amended at Edinburgh in 2000. The study was conducted in accordance with the International Conference on Harmonization Good Clinical Practice (ICH GCP) guidelines. Patient information was anonymized, and any collection of patient data was in compliance with the Data Protection Act 1998. The study underwent full ethical approval by London- Stanmore Research Ethics Committee. REC ref: 16/LO/1325. Written and informed consent was obtained from all participants in the trial. All authors had access to the study data and reviewed and approved the final manuscript.
Study Design
Patients identified to have a confirmed diagnosis of CD (either through histological, endoscopic, and/or radiological evidence) over the age of 18 years with a primary terminal ileal resection were invited to participate in the study. Patients were identified for recruitment within 1-year of their primary resection and were recruited from 2 district general hospitals and one tertiary center. All patients were assessed for diarrhea following their surgery (generally within 4-6 weeks postoperatively) and underwent a 75-selenium homocholic acid taurine (75SeHCAT) scan. Diarrhea was defined as the persistent alteration from the patient’s norm with stool consistency between types 5 and 7 on the Bristol stool chart and increased frequency greater than 4-week duration.30 Patients were excluded from the study if they were: pregnant or breast-feeding; unable to provide written consent; known established bile acid diarrhea; currently or previously treated with bile acid sequestrants; recipients of antibiotics within 4 weeks of the initial trial participation; or had a previous small bowel resection.
As per the UK National Institute for Health and Care Excellence (NICE) guidance, a 75SeHCAT result of <5% was considered as severe bile acid diarrhea, 5%-10% as moderate, 10%-15% as mild and >15% as a negative result.26 Patients with a positive 75SeHCAT result received a therapeutic trial of bile acid sequestrants with colesevelam 625 mg once or twice daily as first-line treatment.
Patients were reviewed in a research clinic 4 and 8 weeks after treatment commencement and assessment of response was made at each review. Patients were required to complete a 7-day stool chart prior to their appointment where daily stool frequency and consistency (as per the Bristol Stool Form Scale) were documented. Stool samples were also collected at each clinic appointment. All patients were consented and booked for a 6-12 month postoperative colonoscopy appointment to assess for disease recurrence, in line with BSG guidelines.31
If patients had a negative 75SeHCAT scan, they provided a single stool sample and were followed up at their postoperative colonoscopy appointment. During this time, all patients received standard therapies and treatments as per normal practice, including immunomodulating and immunosuppressive medications. At their colonoscopy visit, a further stool sample and a 7-day stool diary form were collected prior to bowel prep administration. For all stool collections, an early morning stool collection was advised; however, depending on the patient’s time and ability, a random stool sample was collected from any point in the day and immediately transferred to −80 °C freezer for storage. Additional mucosal samples were taken by biopsy from the following sites in the bowel: neo-terminal ileum, anastomosis, proximal colon, distal colon, and/or rectum. From each site, 2 normal representative biopsies were taken and 2 further biopsies were taken from the most inflamed mucosa. The mucosal biopsies were stored in liquid nitrogen until they could be transferred to a −80 °C freezer.
Clinical response was defined as patients who had improved bowel frequency by >50% from their initial assessment or <3 bowel movements per day. If patients had a partial response (defined as improved bowel frequency but not >50% or reduced bowel frequency but still >3 bowel movements/day), their colesevelam dose was increased at their clinic appointment and reviewed again in 4 weeks’ time. Any side effects of the treatment were documented, as well as a review of their medication history. If patients could not tolerate the medication or no benefit was observed, they were subsequently withdrawn from the study; however, their stool samples collected up to that point were still used for analysis. The endoscopic response was assessed via Rutgeert’s score with a score >i2 indicating disease recurrence.
DNA Extraction and 16S rRNA Amplicon Sequencing
Microbial DNA was extracted from fecal samples according to the manufacturer’s instructions using the commercially available QIAamp Fast DNA Stool Mini Kit (Qiagen, UK). Similarly, microbial DNA was extracted from intestinal biopsies using a modified protocol of Qiagen All Prep DNA mini-Kit (Qiagen, Hilden, Germany). The extracted microbial DNA was then used for 16S ribosomal RNA (rRNA) gene amplification and sequencing to determine the mucosal-adherent microbiota in the mucosal samples as per the Earth Microbiome project protocol.32 Commercially available primers were targeted to the V4 region (515F Parada: GTGYCAGCMGCCGCGGTAA, 806R Apprill: GGACTACNVGGGTWTCTAAT) and the 16S rRNA genes were amplified in triplicate. Each sample was amplified via polymerase chain reaction, with a unique “Earth Microbiome Project” primer (16S Illumina Amplicon Protocol) that had a specific barcode to enable sample identification after sequencing. DNA extraction and 16S rRNA gene polymerase chain reaction were performed via paired-end sequencing (2 × 300 base pairs) using the MiSeq v2 Reagent kit and the Illumina MiSeq system (Illumina, San Diego, USA).
Statistical Analysis
Microbial bioinformatic analysis was performed using the Quantitative Insight into Microbial Ecology 2 pipeline.33 Forward and reverse reads were assessed for quality using qiime demux summaries and trimmed using DADA2 to remove low-quality reads. Rarefaction plots were used to identify sequence sampling depth and α- and β-diversity were then estimated using the rarefied data. High-quality reads were clustered into amplicon sequence variants and classified using the SILVA 16S rRNA gene database using DADA2’s default parameters.
The Shannon diversity index and Faith’s phylogenetic diversity were used to assess α-diversity metrics comparing intra-sample variability. Shannon α-diversity metric accounts for both the richness (total number of species within the community) and the evenness (relative abundance of different species),34 whereas Faith’s phylogenetic diversity is a measure of the biodiversity that incorporates phylogenetic differences between species.35 Statistical analysis for intra-sample comparisons (including pairwise and longitudinal analysis) was assessed using the non-parametric Kruskal–Wallis test. The Bray-Curtis dissimilarity distance matrix was used to assess β-diversity comparing inter-sample variability. Statistical analysis was performed using the permutational multivariate analysis of variance (PERMANOVA).
Comparison of relative abundances of taxa between the different group cohorts was performed using a linear discriminant analysis (LDA) effect size (LEfSe).36 Taxa with an LDA score >2 with a P-value ≤ .05 was considered statistically significant. Corrected q-values to adjust for the false discovery rate were calculated for multiple hypothesis testing between treatment groups and a q-value ≤ 0.05 was considered statistically significant.37
Results
Study Population
A total of 14 patients who underwent a primary ileo-caecal resection participated in the study. Of the 14 patients that provided a pretreatment stool sample, 10 patients provided a subsequent stool sample 4 weeks posttreatment with colesevelam, 9 patients provided a stool sample 8 weeks posttreatment, and 11 patients provided a stool sample 6-12 months following treatment (Figure 1). Matched outcomes were available for 9 patients who provided fecal samples at 4 weeks and 6-12 months with correlating biopsy samples at their colonoscopy.
Study recruitment and flowchart for patients with Crohn’s disease who underwent a primary terminal ileal resection. These patients were assessed for diarrhea 4-6 weeks following their primary resection and those who were symptomatic underwent a 75SeHCAT scan to assess for bile acid diarrhea. Patients with a positive 75SeHCAT (≤15% retention) were given treatment with a bile acid sequestrant. Stool samples were collected during pretreatment and 4 weeks, 8 weeks, and 6-12 months following treatment. Biopsy samples were collected at their 6-12 months colonoscopy appointment. Patients who did not have diarrhea or had a negative 75SeHCAT (>15% retention) were not given treatment and stool samples were collected during pretreatment and 6-12 months later, alongside their biopsy samples at the time of their postoperative colonoscopy appointment.
A total of 11/14 (78.6%) patients were symptomatic with diarrhea and urgency 4-6 weeks postoperatively. In total, 10/11 (90.9%) of patients were diagnosed with severe bile acid diarrhea based on their 75SeHCAT scan and subsequently started on colesevelam. A total of 1/11 (9.1%) patients had a normal 75SeHCAT scan and did not have colesevelam. Of 3/14 (21.4%) patients did not undergo 75SeHCAT investigations as they were asymptomatic and presumed negative for bile acid diarrhea. During the duration of the study, 2/14 (14.3%) patients continued smoking, 2/14 (14.3%) were on thiopurines and 2/14 (14.3%) continued biological medication postoperatively. Of these, 1/14 (7.1%) patients were taking both a thiopurine and biologic. Pre- and postoperative patient demographics are outlined in Tables 1 and 2.
Description of patient demographics, disease characteristics, medication history and surgical procedures for patients with Crohn’s disease prior to undergoing their primary terminal ileal resection.
| Patient | Age | Gender (M/F) | Smoking status (Y/N) | Montreal classification | Thiopurines | Biologics | Previous medications | Elective or emergency surgery | Surgical procedure |
|---|---|---|---|---|---|---|---|---|---|
| A | 23 | M | N | A2L1B1 | Azathioprine 75 mg | Adalimumab ALT weeks | Elective | Right hemicolectomy followed by reversal of ileostomy | |
| B | 44 | F | N | A2L1B1 | Elective | Right hemicolectomy with ileo-colic anastomosis | |||
| C | 33 | F | N | A2L1B1p | Azathioprine 125 mg | Adalimumab ALT weeks | Elective | Right hemicolectomy followed by reversal of ileostomy | |
| D | 29 | M | Y | A2L1B3 | Azathioprine 100 mg | Emergency | Laparotomy and ileocolic resection | ||
| E | 34 | M | N | A2L3B3 | Adalimumab ALT weeks | Emergency | Emergency laparotomy for perforated transverse colon and subtotal colectomy with end ileostomy, followed by elective laparotomy and division of extensive adhesions and ileo-sigmoid anastomosis. | ||
| F | 30 | M | N | A2L3B1 | Elective | Right hemicolectomy and resection of ileocolonic Crohn’s disease | |||
| G | 37 | F | N | A2L1B1 | Elective | Ileocaecal resection and formation of defunctioning ileostomy and mucus fistula, followed by reversal of end ileostomy. | |||
| H | 68 | M | N | A3L1B1 | Elective | Laparotomy and right hemicolectomy with ileocolic anastomosis | |||
| I | 30 | F | N | A2L3B1 | Elective | Laparotomy and right hemicolectomy with ileocolic anastomosis | |||
| J | 78 | F | N | A3L1B2 | Elective | Right hemicolectomy with ileocolic anastomosis | |||
| K | 38 | M | N | A2L1B3 | Elective | Right hemicolectomy and ileostomy followed by reversal of loop ileostomy | |||
| L | 40 | F | N | A2L3B3 | Methotrexate | Vedolizumab 8-weekly | Infliximab Adalimumab | Elective | Right hemicolectomy, resection of entero-colic fistula and repair of sigmoid colon |
| M | 23 | F | N | A2L1B1 | Ustekinumab 8-weekly | Adalimumab | Elective | Limited right hemicolectomy with ileocolonic anastomosis | |
| N | 43 | M | Y | A3L3B1 | Azathioprine 125 mg | Adalimumab ALT weeks | Elective | Limited right hemicolectomy with ileocolonic anastomosis |
| Patient | Age | Gender (M/F) | Smoking status (Y/N) | Montreal classification | Thiopurines | Biologics | Previous medications | Elective or emergency surgery | Surgical procedure |
|---|---|---|---|---|---|---|---|---|---|
| A | 23 | M | N | A2L1B1 | Azathioprine 75 mg | Adalimumab ALT weeks | Elective | Right hemicolectomy followed by reversal of ileostomy | |
| B | 44 | F | N | A2L1B1 | Elective | Right hemicolectomy with ileo-colic anastomosis | |||
| C | 33 | F | N | A2L1B1p | Azathioprine 125 mg | Adalimumab ALT weeks | Elective | Right hemicolectomy followed by reversal of ileostomy | |
| D | 29 | M | Y | A2L1B3 | Azathioprine 100 mg | Emergency | Laparotomy and ileocolic resection | ||
| E | 34 | M | N | A2L3B3 | Adalimumab ALT weeks | Emergency | Emergency laparotomy for perforated transverse colon and subtotal colectomy with end ileostomy, followed by elective laparotomy and division of extensive adhesions and ileo-sigmoid anastomosis. | ||
| F | 30 | M | N | A2L3B1 | Elective | Right hemicolectomy and resection of ileocolonic Crohn’s disease | |||
| G | 37 | F | N | A2L1B1 | Elective | Ileocaecal resection and formation of defunctioning ileostomy and mucus fistula, followed by reversal of end ileostomy. | |||
| H | 68 | M | N | A3L1B1 | Elective | Laparotomy and right hemicolectomy with ileocolic anastomosis | |||
| I | 30 | F | N | A2L3B1 | Elective | Laparotomy and right hemicolectomy with ileocolic anastomosis | |||
| J | 78 | F | N | A3L1B2 | Elective | Right hemicolectomy with ileocolic anastomosis | |||
| K | 38 | M | N | A2L1B3 | Elective | Right hemicolectomy and ileostomy followed by reversal of loop ileostomy | |||
| L | 40 | F | N | A2L3B3 | Methotrexate | Vedolizumab 8-weekly | Infliximab Adalimumab | Elective | Right hemicolectomy, resection of entero-colic fistula and repair of sigmoid colon |
| M | 23 | F | N | A2L1B1 | Ustekinumab 8-weekly | Adalimumab | Elective | Limited right hemicolectomy with ileocolonic anastomosis | |
| N | 43 | M | Y | A3L3B1 | Azathioprine 125 mg | Adalimumab ALT weeks | Elective | Limited right hemicolectomy with ileocolonic anastomosis |
Abbreviation: ALT, alternate.
Description of patient demographics, disease characteristics, medication history and surgical procedures for patients with Crohn’s disease prior to undergoing their primary terminal ileal resection.
| Patient | Age | Gender (M/F) | Smoking status (Y/N) | Montreal classification | Thiopurines | Biologics | Previous medications | Elective or emergency surgery | Surgical procedure |
|---|---|---|---|---|---|---|---|---|---|
| A | 23 | M | N | A2L1B1 | Azathioprine 75 mg | Adalimumab ALT weeks | Elective | Right hemicolectomy followed by reversal of ileostomy | |
| B | 44 | F | N | A2L1B1 | Elective | Right hemicolectomy with ileo-colic anastomosis | |||
| C | 33 | F | N | A2L1B1p | Azathioprine 125 mg | Adalimumab ALT weeks | Elective | Right hemicolectomy followed by reversal of ileostomy | |
| D | 29 | M | Y | A2L1B3 | Azathioprine 100 mg | Emergency | Laparotomy and ileocolic resection | ||
| E | 34 | M | N | A2L3B3 | Adalimumab ALT weeks | Emergency | Emergency laparotomy for perforated transverse colon and subtotal colectomy with end ileostomy, followed by elective laparotomy and division of extensive adhesions and ileo-sigmoid anastomosis. | ||
| F | 30 | M | N | A2L3B1 | Elective | Right hemicolectomy and resection of ileocolonic Crohn’s disease | |||
| G | 37 | F | N | A2L1B1 | Elective | Ileocaecal resection and formation of defunctioning ileostomy and mucus fistula, followed by reversal of end ileostomy. | |||
| H | 68 | M | N | A3L1B1 | Elective | Laparotomy and right hemicolectomy with ileocolic anastomosis | |||
| I | 30 | F | N | A2L3B1 | Elective | Laparotomy and right hemicolectomy with ileocolic anastomosis | |||
| J | 78 | F | N | A3L1B2 | Elective | Right hemicolectomy with ileocolic anastomosis | |||
| K | 38 | M | N | A2L1B3 | Elective | Right hemicolectomy and ileostomy followed by reversal of loop ileostomy | |||
| L | 40 | F | N | A2L3B3 | Methotrexate | Vedolizumab 8-weekly | Infliximab Adalimumab | Elective | Right hemicolectomy, resection of entero-colic fistula and repair of sigmoid colon |
| M | 23 | F | N | A2L1B1 | Ustekinumab 8-weekly | Adalimumab | Elective | Limited right hemicolectomy with ileocolonic anastomosis | |
| N | 43 | M | Y | A3L3B1 | Azathioprine 125 mg | Adalimumab ALT weeks | Elective | Limited right hemicolectomy with ileocolonic anastomosis |
| Patient | Age | Gender (M/F) | Smoking status (Y/N) | Montreal classification | Thiopurines | Biologics | Previous medications | Elective or emergency surgery | Surgical procedure |
|---|---|---|---|---|---|---|---|---|---|
| A | 23 | M | N | A2L1B1 | Azathioprine 75 mg | Adalimumab ALT weeks | Elective | Right hemicolectomy followed by reversal of ileostomy | |
| B | 44 | F | N | A2L1B1 | Elective | Right hemicolectomy with ileo-colic anastomosis | |||
| C | 33 | F | N | A2L1B1p | Azathioprine 125 mg | Adalimumab ALT weeks | Elective | Right hemicolectomy followed by reversal of ileostomy | |
| D | 29 | M | Y | A2L1B3 | Azathioprine 100 mg | Emergency | Laparotomy and ileocolic resection | ||
| E | 34 | M | N | A2L3B3 | Adalimumab ALT weeks | Emergency | Emergency laparotomy for perforated transverse colon and subtotal colectomy with end ileostomy, followed by elective laparotomy and division of extensive adhesions and ileo-sigmoid anastomosis. | ||
| F | 30 | M | N | A2L3B1 | Elective | Right hemicolectomy and resection of ileocolonic Crohn’s disease | |||
| G | 37 | F | N | A2L1B1 | Elective | Ileocaecal resection and formation of defunctioning ileostomy and mucus fistula, followed by reversal of end ileostomy. | |||
| H | 68 | M | N | A3L1B1 | Elective | Laparotomy and right hemicolectomy with ileocolic anastomosis | |||
| I | 30 | F | N | A2L3B1 | Elective | Laparotomy and right hemicolectomy with ileocolic anastomosis | |||
| J | 78 | F | N | A3L1B2 | Elective | Right hemicolectomy with ileocolic anastomosis | |||
| K | 38 | M | N | A2L1B3 | Elective | Right hemicolectomy and ileostomy followed by reversal of loop ileostomy | |||
| L | 40 | F | N | A2L3B3 | Methotrexate | Vedolizumab 8-weekly | Infliximab Adalimumab | Elective | Right hemicolectomy, resection of entero-colic fistula and repair of sigmoid colon |
| M | 23 | F | N | A2L1B1 | Ustekinumab 8-weekly | Adalimumab | Elective | Limited right hemicolectomy with ileocolonic anastomosis | |
| N | 43 | M | Y | A3L3B1 | Azathioprine 125 mg | Adalimumab ALT weeks | Elective | Limited right hemicolectomy with ileocolonic anastomosis |
Abbreviation: ALT, alternate.
Patient demographics following their primary terminal ileal resection. Patients were diagnosed with bile acid diarrhea if they had a 75SeHCAT ≤ 15% retention and were started on bile acid sequestrants (colesevelam). Response to colesevelam was defined as patients who had improved bowel frequency by >50% from their initial assessment or <3 bowel movements per day. All patients received standard therapy and treatment, as per normal practice, including immunomodulating and immunosuppressive therapy. A Rutgeert’s score of <i2 was considered as endoscopic remission, whereas a score ≥i2 was considered to be disease relapse.
| Patient | 75SeHCAT (Y/N) | 75SeHCAT result | Colesevelam (Y/N) | Response to colesevelam (Y/N) | Smoking status (Y/N) | Thiopurines | Biologics | Colonoscopy Rutgeert’s score | Biopsy results | MRE results |
|---|---|---|---|---|---|---|---|---|---|---|
| A | Y | 1.5% | Y | Y | N | I1 | No evidence of active disease | |||
| B | Y | 0.9% | Y | Y | N | I1 | ||||
| C | Y | 1.8% | Y | Y | N | Azathioprine 125 mg OD | Adalimumab ALT weeks | I0 | Normal mucosa | |
| D | Y | 0.4% | Y | Y | Y | N/A | No evidence of active disease | |||
| E | Y | 2.2% | Y | Y | N | Adalimumab ALT weeks | I1 | Mildly active chronic inflammation in neo-TI | ||
| F | Y | 0.9% | Y | Y | N | Azathioprine 150/175 mg | I2 | Mildly active chronic inflammation at anastomosis | ||
| G | N | N | N | I3 | Focal mild active chronic inflammation at anastomosis | Short segment (3cm) of mild inflammation changes in neo-TI | ||||
| H* | N | N | N | |||||||
| I | Y | 17.4% | N | N | I0 | Quiescent disease | ||||
| J | Y | 1.8% | Y | Y | N | I0 | Normal mucosa | |||
| K** | N | N | N | Limited views | Acute on chronic neo-TI (10cm) | |||||
| L | Y | 0.6% | Y | Y | N | I0 | ||||
| M | Y | 0.7% | Y | Y | N | Ustekinumab 8-weekly | I1 | Quiescent disease | ||
| N | Y | 2.2% | Y | Y | Y | I1 | Normal mucosa |
| Patient | 75SeHCAT (Y/N) | 75SeHCAT result | Colesevelam (Y/N) | Response to colesevelam (Y/N) | Smoking status (Y/N) | Thiopurines | Biologics | Colonoscopy Rutgeert’s score | Biopsy results | MRE results |
|---|---|---|---|---|---|---|---|---|---|---|
| A | Y | 1.5% | Y | Y | N | I1 | No evidence of active disease | |||
| B | Y | 0.9% | Y | Y | N | I1 | ||||
| C | Y | 1.8% | Y | Y | N | Azathioprine 125 mg OD | Adalimumab ALT weeks | I0 | Normal mucosa | |
| D | Y | 0.4% | Y | Y | Y | N/A | No evidence of active disease | |||
| E | Y | 2.2% | Y | Y | N | Adalimumab ALT weeks | I1 | Mildly active chronic inflammation in neo-TI | ||
| F | Y | 0.9% | Y | Y | N | Azathioprine 150/175 mg | I2 | Mildly active chronic inflammation at anastomosis | ||
| G | N | N | N | I3 | Focal mild active chronic inflammation at anastomosis | Short segment (3cm) of mild inflammation changes in neo-TI | ||||
| H* | N | N | N | |||||||
| I | Y | 17.4% | N | N | I0 | Quiescent disease | ||||
| J | Y | 1.8% | Y | Y | N | I0 | Normal mucosa | |||
| K** | N | N | N | Limited views | Acute on chronic neo-TI (10cm) | |||||
| L | Y | 0.6% | Y | Y | N | I0 | ||||
| M | Y | 0.7% | Y | Y | N | Ustekinumab 8-weekly | I1 | Quiescent disease | ||
| N | Y | 2.2% | Y | Y | Y | I1 | Normal mucosa |
Abbreviations: BAS, bile acid sequestrant therapy; MRE, magnetic resonance enterography; TI, terminal ileum; ALT, alternate. *Patient refused 6-12 months colonoscopy check and MRE. **Colonoscopy results had limited views due to poor bowel prep and went onto have an MRE instead.
Patient demographics following their primary terminal ileal resection. Patients were diagnosed with bile acid diarrhea if they had a 75SeHCAT ≤ 15% retention and were started on bile acid sequestrants (colesevelam). Response to colesevelam was defined as patients who had improved bowel frequency by >50% from their initial assessment or <3 bowel movements per day. All patients received standard therapy and treatment, as per normal practice, including immunomodulating and immunosuppressive therapy. A Rutgeert’s score of <i2 was considered as endoscopic remission, whereas a score ≥i2 was considered to be disease relapse.
| Patient | 75SeHCAT (Y/N) | 75SeHCAT result | Colesevelam (Y/N) | Response to colesevelam (Y/N) | Smoking status (Y/N) | Thiopurines | Biologics | Colonoscopy Rutgeert’s score | Biopsy results | MRE results |
|---|---|---|---|---|---|---|---|---|---|---|
| A | Y | 1.5% | Y | Y | N | I1 | No evidence of active disease | |||
| B | Y | 0.9% | Y | Y | N | I1 | ||||
| C | Y | 1.8% | Y | Y | N | Azathioprine 125 mg OD | Adalimumab ALT weeks | I0 | Normal mucosa | |
| D | Y | 0.4% | Y | Y | Y | N/A | No evidence of active disease | |||
| E | Y | 2.2% | Y | Y | N | Adalimumab ALT weeks | I1 | Mildly active chronic inflammation in neo-TI | ||
| F | Y | 0.9% | Y | Y | N | Azathioprine 150/175 mg | I2 | Mildly active chronic inflammation at anastomosis | ||
| G | N | N | N | I3 | Focal mild active chronic inflammation at anastomosis | Short segment (3cm) of mild inflammation changes in neo-TI | ||||
| H* | N | N | N | |||||||
| I | Y | 17.4% | N | N | I0 | Quiescent disease | ||||
| J | Y | 1.8% | Y | Y | N | I0 | Normal mucosa | |||
| K** | N | N | N | Limited views | Acute on chronic neo-TI (10cm) | |||||
| L | Y | 0.6% | Y | Y | N | I0 | ||||
| M | Y | 0.7% | Y | Y | N | Ustekinumab 8-weekly | I1 | Quiescent disease | ||
| N | Y | 2.2% | Y | Y | Y | I1 | Normal mucosa |
| Patient | 75SeHCAT (Y/N) | 75SeHCAT result | Colesevelam (Y/N) | Response to colesevelam (Y/N) | Smoking status (Y/N) | Thiopurines | Biologics | Colonoscopy Rutgeert’s score | Biopsy results | MRE results |
|---|---|---|---|---|---|---|---|---|---|---|
| A | Y | 1.5% | Y | Y | N | I1 | No evidence of active disease | |||
| B | Y | 0.9% | Y | Y | N | I1 | ||||
| C | Y | 1.8% | Y | Y | N | Azathioprine 125 mg OD | Adalimumab ALT weeks | I0 | Normal mucosa | |
| D | Y | 0.4% | Y | Y | Y | N/A | No evidence of active disease | |||
| E | Y | 2.2% | Y | Y | N | Adalimumab ALT weeks | I1 | Mildly active chronic inflammation in neo-TI | ||
| F | Y | 0.9% | Y | Y | N | Azathioprine 150/175 mg | I2 | Mildly active chronic inflammation at anastomosis | ||
| G | N | N | N | I3 | Focal mild active chronic inflammation at anastomosis | Short segment (3cm) of mild inflammation changes in neo-TI | ||||
| H* | N | N | N | |||||||
| I | Y | 17.4% | N | N | I0 | Quiescent disease | ||||
| J | Y | 1.8% | Y | Y | N | I0 | Normal mucosa | |||
| K** | N | N | N | Limited views | Acute on chronic neo-TI (10cm) | |||||
| L | Y | 0.6% | Y | Y | N | I0 | ||||
| M | Y | 0.7% | Y | Y | N | Ustekinumab 8-weekly | I1 | Quiescent disease | ||
| N | Y | 2.2% | Y | Y | Y | I1 | Normal mucosa |
Abbreviations: BAS, bile acid sequestrant therapy; MRE, magnetic resonance enterography; TI, terminal ileum; ALT, alternate. *Patient refused 6-12 months colonoscopy check and MRE. **Colonoscopy results had limited views due to poor bowel prep and went onto have an MRE instead.
Twelve patients underwent a 6-12 month postoperative colonoscopy although mucosal biopsies were obtained in only 9 patients. Two patients declined colonoscopy assessment of which 1 patient chose to have radiological imaging with magnetic resonance imaging of the small bowel (MRE) and the other patient declined all investigations as was asymptomatic. One patient had limited views at colonoscopy and so underwent an MRE as well.
Quality control
A total of 44 fecal samples and 44 mucosal biopsies were sequenced from 14 patients. Following quality control, 11.7 million reads were obtained with an average of 132 679 reads/sample (standard deviation of 71 210 reads/sample). A subsampling depth of 5044 reads/sample was chosen following rarefaction.
Treatment response
Of the 10 patients who were given colesevelam, 9 (90%) had a complete clinical response to treatment. One patient had a partial response and continued taking the colesevelam until the colonoscopy appointment. At their 6-12 months review, 9/10 (90%) were found to be in disease remission, of which 8 showed endoscopic remission (Rutgeert’s score of either i0 or i1) and 1 patient demonstrated radiological remission (MRE). While 7 of the 14 patients were either on immunomodulating and/or immunosuppressive therapy preoperatively, 4 of these patients restarted medication postoperatively. Only 1/10 (10%) of those patients on colesevelam had a Rutgeert’s score of i2, indicating endoscopic recurrence. This patient was not on immunosuppressive or immunomodulating therapy and did not have any additional risk factors such as smoking.
Of the 4 patients who did not receive colesevelam, none of them were on additional medication for their CD or were considered high-risk patients for recurrence. Fifty percent of patients (2/4) displayed evidence of recurrence (i3 on endoscopy and active disease seen on MRE). Twenty-five percent of patients (1/4) had an endoscopic Rutgeert’s score of i0 and the remaining patient declined further investigations as was asymptomatic.
Bacterial diversity in patients with Crohn’s disease treated with bile acid sequestrants
Eleven of the 14 patients submitted paired baseline, pretreatment stool samples, and 6-12 months posttreatment with colesevelam stool samples. Comparison of α-diversity metrics (Shannon Diversity, Faith’s phylogenetic diversity) did not demonstrate any differences in postoperative patients between pretreatment and 6-12 months posttreatment (Figure 2A and 2B, P = NS). Consistent with this, the Bray-Curtis distance matrix assessed β-diversity between the 2 groups showed no differences in bacterial community clusters (Figure 2C, P = NS).
Bacterial diversity and microbial composition in patients with postoperative Crohn’s disease who were also diagnosed with bile acid diarrhea at baseline (pretreatment) and following 6-12 months of treatment with a bile acid sequestrant (colesevelam). α-diversity metrics were analyzed using (A) Shannon diversity and (B) Faith’s phylogenetic diversity. (C) β-diversity metric was analyzed with Bray Curtis distance matrix. (D) LEfSE histogram of LDA scores for differentially abundant bacterial taxa in the pretreatment cohort. Taxa with an LDA > 2 with a P-value ≤ .05 was considered statistically significant. Boxplots represent the mean and standard error of the mean.
The LDA effect size (LEfSe) was used to rank the greatest differences in abundance genera between the pre- and posttreatment cohort (Figure 2D). Significant changes were only seen in the pretreatment group with an enrichment in the Pseudomonadales and Christensenellales orders; Muribaculaceae and Christensenellaceae families; and Muribaculaceae, UCG 004, Holdemanella, Subdoligranulum, and CAG 352 genera. One bacterial species was unidentifiable.
Paired individual bacterial changes are visualized between pre- and posttreatment groups in Figure 3. Nine patients were treated with colesevelam and 2 patients were not given colesevelam. Pretreatment, the 3 most abundant bacterial classes in all patients were Bacteroidia, Clostridia, and Gammaproteobacteria. Following 6 months of treatment, out of the 9 patients on colesevelam, 5/9 (55.6%) had a reduction in Bacteroidia, 9/9 (100%) had an increase in Clostridia, and 7/9 (77.8%) had a reduction in Gammaproteobacteria. Of the 2 patients not given colesevelam, One of the two patients had a reduction in Bacteroidia, an increase in Clostridia, and a reduction in Gammaproteobacteria. Paired student’s tests were not significant.
Composition of the microbiome in paired patients with postoperative Crohn’s disease at baseline (pre) and 6-12 months postoperatively. Patients A, B, C, E, J, M, and N were diagnosed with bile acid diarrhea and were given treatment with a bile acid sequestrant (colesevelam) for up to 12 months. Patients G and H (denoted in *) were not diagnosed with bile acid diarrhea and were not given treatment with a bile acid sequestrant. The relative abundance frequency ratios of each microorganism group are shown in different colors.
Changes to microbiome based on endoscopic severity
We explored whether there were changes to bacterial diversity in patients who demonstrated endoscopic recurrence (Rutgeert’s score ≥ i2) compared to those patients in remission (Rutgeert’s score ≤ i1) at their 6-12 months postoperative colonoscopy appointment. There were no differences in α-diversity demonstrated with either Shannon's diversity and Faith’s phylogenetic diversity (Figure 4A and 4B; P = NS). There were no significant differences in specific bacterial taxa between Rutgeert’s scores i0 to i3.
Comparison of bacterial diversity in patients with Crohn’s disease following a primary terminal ileal resection at their 6-12 months post-operative colonoscopy appointment. Endoscopic recurrence was determined by Rutgeert’s score where endoscopic remission was considered a score of i0 and i1 and endoscopic recurrence was based on a score of i2 and i3. α-diversity metrics were analyzed using (A) Shannon diversity; P = nonsignificant and (B) Faith’s phylogenetic diversity; P = nonsignificant. Boxplots represent the mean and standard error of the mean.
Discussion
This pilot study is the first to investigate the effects of colesevelam on the postoperative recurrence risk in patients with CD. Firstly, we demonstrated that 90% of patients on colesevelam, but only 50% without colesevelam, were in endoscopic remission 6-12 months postoperatively. Despite our study being small, this is still a noteworthy finding considering previous studies have shown that the postoperative recurrence rate without treatment is between 65% and 90% within 12 months and 80%-100% within 3 years from the primary surgery.38-40 This risk will increase by 25% every year and eventually more than 50% of patients will need a second operation within 10 years from their initial resection.41,42 To demonstrate that only 10% of our patients on colesevelam, while 50% without colesevelam had recurrent disease within their first year of surgery, highlights the need for greater research in this area.
Risk factors for recurrence include smoking, perianal disease, prior intestinal surgery, extended small bowel resection of >100 cm, and penetrating disease behavior including fistulisation, perforation, and intra-abdominal or retroperitoneal abscesses.43-47 It is recommended that high-risk patients should be considered for postoperative prophylactic therapy with advanced therapies to prevent the risk of early relapse.38,48 As this was a pilot study, we followed national BSG guidelines on managing postoperative CD so that patients were not negatively impacted by participating in our trial.31 Out of the 14 patients, 3 were on prophylactic treatment with either immunomodulating or immunosuppressive therapy. Two of these patients were admitted to hospital with a perforated bowel and had emergency surgery. One of them continued to smoke following hospital discharge. The third patient had active perianal fistulising disease with a seton in situ. Interestingly, the 2 patients who were not taking colesevelam who showed recurrence on their colonoscopy did not have any high-risk features and were not on prophylactic treatment.
In the examination of the microbiome, there was no significant change in α/β-diversity before and after 6-12 months of treatment with colesevelam. In the pretreatment samples, the LEfSE histogram demonstrated a bacterial abundance of the Christensenellaceae order and family, and the Holdemanella and Subdoligranulum genera, all of which have been associated with a healthy metabolic status.49-56 A possible explanation for our findings is that the baseline stool sample was collected 4-6 weeks postoperatively after the diseased bowel segment had been removed and it is well-known that surgery itself can alter the microbiome.57-59 Mondot et al. demonstrated that while patients with an ileocolonic resection did not have a change in microbial diversity 6 months postoperatively, there was an alteration in microbial community structure in those patients with endoscopic recurrence.59 The majority of our patients were in endoscopic remission at the time of their 6-12 months stool collection so changes may not have been as noticeable if stool samples were also collected prior to their surgery.
When the patients’ individual microbial composition was directly compared pre- and posttreatment, 100% showed an increased relative frequency of Clostridia and 78% showed a reduction in Gammaproteobacteria following treatment with colesevelam. Clostridia, also part of the Firmicutes phylum, are key sources of the short-chain fatty acid butyrate, which is an important energy source of intestinal epithelial cells and can promote wound healing.60,61 In mice models, butyrate has been shown to exert anti-inflammatory and pro-intestinal barrier properties.62,63 In contrast, increased levels of Gammaproteobacteria have been implicated in the pathogenesis of CD.58,64-66 Recently, Gonzalez et al. demonstrated that ileal CD is largely distinguished by increased affinity for a bile acid-rich environment along with concomitant shifts in taxa with known associations with bile acids, which includes Gammaproteobacteria.67 Thus, our data suggests that treatment with colesevelam may have a role in altering the microbiome into a healthier status to maintain remission states in CD.
Previous studies have explored changes to the fecal microbiota composition in postoperative CD, but few studies have correlated mucosal samples with endoscopically confirmed recurrence.68,69 Our biopsy sample results did not demonstrate any significant changes in bacterial diversity between patients in remission and those in recurrence. This is in contrast to Sokol et al. who demonstrated that patients with endoscopic recurrence exhibited strong changes in their ileal mucosa-associated microbiota compared with patients without recurrence.70 The difference in results may be attributed to sample size considering they had 288 samples for comparison and we had 88 samples. Furthermore, their mucosal samples were taken at the ileum where inflammation was seen on endoscopy, whereas we combined our mucosal samples within the colon and neo-terminal ileum rather than simply looking at the affected changes at the anastomosis site.
There are several limitations to this study. Firstly, our results may have been skewed by using postoperative fecal samples as baseline rather than preoperative fecal and mucosal samples. Despite this, we still demonstrated important microbial composition findings in the immediate postoperative period. Secondly, despite the recent advances in 16S rRNA sequencing, there are multiple factors that can affect the microbiome including but not limited to medications, smoking status, age, diet, and gender.66,71,72 While it may be difficult to control for each of these factors, we optimized the standardization of fecal and mucosal collection, storage, transportation, and DNA extraction. Furthermore, the use of 16S rRNA amplicon sequence variants does not provide functional information which is needed to better understand host-microbe-metabolome interactions relevant to states of health and disease. Thus, the study of metagenomics, meta-transcriptomics, and metabolomics would be more informative in combination with patients’ clinical course, medications, histology, immune indices, and mucosal gene expression. Finally, we acknowledge our sample size was small, but this was a pilot mechanistic study exploring for the signal of efficacy. This study was not adequately powered to determine statistical significance and therefore we are only able to provide probable associations rather than cause and/or consequence and the latter is needed to change clinical management.
This study provided crucial preliminary data that demonstrates the need for a larger-scale quantitative study. Indeed, this is the first study to explore an exciting novel therapy in patients with postoperative CD where the safety profile of the medication is already known and more importantly, does not carry the same side effects and long-term complications as the current immunomodulators and advanced therapies. In conclusion, we were able to demonstrate that the bile acid-sequestrant, colesevelam, may have the potential ability to prevent clinical and endoscopic recurrence by manipulating the microbiome. Whether this is achieved through bile acid modulation remains unknown and warrants further study.
Acknowledgments
None.
Author Contributions
A.K. led the study, recruited, collected data, and analyzed the results. She wrote the first draft of the manuscript. N.Q., M.E., and A.B. aided with data analysis. M.J., H.S., J.B., and A.F. helped with study recruitment. J.M., H.O.A., and M.J.B. conceptualized and designed the study. A.K., N.Q., A.B., H.O.A., J.P.S., and M.J.B. were involved with critical revisions of the manuscript. All authors read and approved the final manuscript.
Funding
The research department of M.J.B. received project funding from Bowel and Cancer Research for part of this work; The research department of M.J.B. received project funding from an unrestricted grant from Tillotts Pharma for part of this work. A.D.B. receives salary support from an M.R.C. Senior Fellow award (Reference MR/X006433/1).
Conflicts of Interest
M.J.B. has received grants and travel expenses from Vifor International and Tillotts Pharma, outside of the submitted work. The research department of M.J.B. also received funding from Tillotts Pharma to support part of the described work. H.S. has received travel and conference expenses from Tillotts Pharma, Norgine, M.S.D., Abbvie and Janssen outside of the submitted work. J.P.S. has received speaker fees for Abbvie, Takeda, and Janssen outside of the submitted work. The remaining authors declare no competing interests. A.K. has received speaker fees for Abbvie and Ferring and has been on the advisory board for Abbvie and Dr Falk, outside of the submitted work.
Data Availability
The datasets presented in this study have been deposited in the NCBI repository, accession number PRJNA941862.
