Short chain fatty acids (butyrate) for induction of remission in ulcerative colitis

Abstract

This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:

The primary objective is to assess the efficacy and safety of SCFAs for induction of remission in active UC. 

Background

Description of the condition

Ulcerative colitis (UC) is a chronic inflammatory disease of the colon, manifested by periods of remissions and exacerbations with bloody diarrhea, abdominal pain, urgency, and tenesmus. The disease can be anatomically classified as proctitis, left side colitis (distal to the splenic flexure) or extensive colitis (Silverberg 2005). Although topical enema treatment with 5‐ASA or corticosteroids is effective in limited distal disease, it is also effective in extensive colitis when used in conjunction with systemic treatments. Adult onset disease is typically confined to the left colon whereas extensive disease is twice as common in children (Van Limbergen 2008).

Aminosalicylates (5‐ASA), corticosteroids, and TNF‐alpha antagonists are the most commonly used therapies for induction of remission in UC. Thiopurines, 5‐ASA and infliximab have been shown to effectively maintain remission in the clinical trial setting (Rutgeerts 2005; Baumgart 2008). Whereas a top‐down approach is increasingly considered in Crohn's disease, the current treatment paradigm to maintain remission in UC is still based on a "step up model", beginning with 5‐ASA preparations. Azathioprine, followed by an anti‐TNF therapy, may be considered in 5‐ASA resistant cases. Surgery remains a valid option in refractory cases and in those with complications such as perforation, uncontrolled bleeding and neoplasia.

Description of the intervention

SHORT CHAIN FATTY ACIDS

Short chain fatty acids (SCFAs, organic acids) are the end stage product of anaerobic fermentation of undigested carbohydrates, oligosaccharides, oligopeptides and degradable fibre in the large bowel, yielding acetate, propionate and butyrate. The latter three are composed of two‐, three‐, and four‐ carbon short chain fatty acids, respectively. Short chain fatty acids provide an important source of energy for colonocytes via an intracellular ß‐ oxidation pathway (Roediger 1982a). Pennington et al first showed that the oxidation pathway ends in ketone bodies and Co₂ (Peninngton 1957; Pennington 1956; Hird 1966). The ability to oxidize butyrate is one of the metabolic differences between the mucosa of the large and small intestine (Roediger 1980a). Indeed, over 70% of the oxygen consumption of colonocytes is associated with butyrate oxidation (Roediger 1980a).

How the intervention might work

SHORT CHAIN FATTY ACIDS AND UC

Roediger et al showed that colonocytes from patients with UC had a significantly reduced ability to oxidize butyrate, proposing that inflammation in UC may be associated with an energy deficiency state (Roediger 1980b). Regardless of the absorption and utilization of SCFAs, fecal SCFA levels were found to be similar in UC patients and healthy controls (Verina 1988; Roediger 1982b; Treem 1994). The finding that the mucosa of the distal colon depends primarily on butyrate while the proximal colon depends primarily on glucose and glutamine, may account for the fact that the inflammation in UC is confined to the colon, more commonly in its distal part Roediger 1980a; however, this remains merely a speculation. Recently, defective butyrate oxidation was found in pouchitis, similar to that seen in active UC but not in quiescence disease (De Preter 2009). These findings suggest that the metabolic defect may be the effect rather than the cause of the inflammation. Similar conclusions were reached in both human and in animal studies (Finnie 1993; Simpson 2000).

Other studies suggest that SCFAs reduce inflammation via a prebiotic effect (Guarner 2005). Evidence from basic science, animal models and clinical research show the importance of the gut microbiome in IBD pathogenesis. Probiotics, especially VSL#3, may be effective for colonic inflammation in the clinical trial setting (Seksik 2008; Tursi 2010).  It is thus likely that the change in the microflora triggered by SCFAs may exert an immunomodulatory effect. Indeed, Lactobacillus casei can downregulate the release of TNF‐alpha from inflamed tissue (Borruel 2002) with an increased number of activated T‐lymphocytes which undergo apoptosis (Carol 2003).

Regardless of whether the defect in oxidizing SCFAs is the cause or the effect, supraphysiologic luminal SCFA concentration may overcome this partial metabolic defect, thus acting as an anti‐inflammatory agent (Kim 1998; Avivi‐Green 2002; Lührs 2002; Segain Gut 2000). SCFAs were found to have a direct anti‐inflammatory effect in a rat model wherein butyrate had an inhibitory effect on lymphocyte proliferation and on the expression of IL‐2 and INF‐ gamma Cavaglieri 2003. In humans, butyrate inhibits the strong pro‐inflammatory transcription factor, NF‐κB, in macrophages isolated from patient with UC (Lührs 2002).

Why it is important to do this review

Following the biological rationale, SCFAs have been the subject of multiple studies evaluating its effectiveness for distal colitis (Breuer 1991). In 1989, Harig and Cloudia showed for the first time that topical SCFAs may be effective for treating diversion colitis (Harig 1989), but other studies did not replicate this finding (Guillemot 1991; Schouten 1991). Several controlled and uncontrolled studies have assessed the role of SCFAs in the treatment paradigm for UC, with inconsistent results (Breuer 1991; Steinhart 1994; Vernia 1995; Patz 1996; Scheppach 1992; Senagore 1992; Steinhart 1996; Kim 1998; Sandborn 1998). Therefore, a systematic review of the literature is planned to assess the role of SCFAs for induction of remission in UC.

Objectives

The primary objective is to assess the efficacy and safety of SCFAs for induction of remission in active UC. 

Methods

Criteria for considering studies for this review

Types of studies

Randomized controlled trials (RCT) published in any language comparing SCFAs versus placebo or standard medical therapy for the treatment of active UC will be considered for inclusion.  Studies published as an abstract form will be included if enough data are provided to ascertain the study methodology and quality.

Types of participants

Studies with patients who have active UC at the time of recruitment will be included. Participants will not be restricted by extent of disease. Both pediatric and adult studies will be included without age restriction.

Types of interventions

Trials, in which SCFAs were administered to induce remission in UC, will be included. Any type of oral or rectally administered SCFA (i.e. butyrate, propionate, and acetate) at any dose or frequency will be considered for inclusion. The comparison group may be a placebo or standard medical therapy, such as 5‐ASA or corticosteroids, administered rectally or systemically. Co‐interventions will be allowed as long as they are balanced between the study groups.

Types of outcome measures

Primary outcomes

The primary outcome of this review is clinical remission within 3 months of intervention as defined by the studies and expressed as a percentage of the number of patients randomized (intention to treat analysis).  Since there are multiple disease activity indices in UC and none is consistently employed Turner 2009, all outcome measures will be accepted for the primary and secondary outcomes of this review as long as they are explicitly presented and well defined. This factor will be explored as a source of heterogeneity. 

Secondary outcomes

Secondary outcomes will include:

1. Clinical improvement within 3 months, as defined by the studies;

2. Endoscopic remission within 3 months as defined by the studies;

3. Endoscopic improvement within 3 months as defined by the studies;

4. Improvement in quality of life (QOL), when applicable (using patient global assessment and scores of generic QOL scales (i.e. IMPACT in children and IBD‐Q in adults); and

5. Rate of adverse events.

Search methods for identification of studies

Electronic searches

An on‐line database literature search will be performed for human studies, without language restrictions, using the following databases: EMBASE (from 1980 to present), Healthstar on OVID (from 1966 to present) , as well as PubMed (from 1950 to present), Cochrane Central Register of Controlled Trials (The Cochrane Library, 2010) and ACP journal club. On‐going trials will be identified from the registry link: www.clinicaltrials.gov. 

For the PubMed search, the following search strategy will be used. The strategy will be adapted for the other databases as appropriate (lines 1‐5, UC; lines 6‐10, SCFA; Lines 11‐13, incorporation):

1. Colitis, Ulcerative [Mesh] OR "ulcerative colitis" [TW] OR Colitis [TW] OR UC [TW].

2. Sigmoiditis [TW]

3. proctitis [Mesh] OR proctitis [TW].

4. Inflamatory Bowel Diseases [Mesh] OR" inflammatory bowel diseases"[TW] OR IBD [TW].

5. OR/ 1‐4

6. Fatty Acids, Volatile [Mesh] OR "short chain fatty acids" [TW] OR SCFA [TW].

7. Butyrate [Mesh] OR Butyric Acids [Mesh] OR butyrate [TW] OR butyrates [TW] OR "sodium butyrate [TW] OR "butyric acid" [TW].

8. Propionates [Mesh] OR Propionic Acids [Mesh] OR propionate [TW] OR "propionic acid" [TW].

9.  Acetic Acids [Mesh] OR acetate [TW] OR "acetic acid" [TW].

10. OR/7‐9

11. 5 and 10

12. (humans NOT animals)

13. 12 AND 11

The search strategy will not be restricted to randomized controlled studies, since the overall literature about UC and short chain fatty acids is limited. All abstracts will be manually reviewed for eligibility.

Searching other resources

Additional citations will be identified by manually searching references from the screened studies and from relevant review articles. Unpublished studies will be sought by contacting experts in the field and from the registry www.clinicaltrials.gov. Proceedings of major international gastroenterology conferences during the previous 3 years as published in the journals Gastroenterology, Inflammatory Bowel Diseases, Journal of Pediatric Gastroenterology and Nutrition, American Journal of Gastroenterology and Gut will be manually screened for eligibility by one author (YK).

Data collection and analysis

Selection of studies

Abstracts of all studies identified by the above search strategy will be screened for eligibility. Full text articles will be retrieved if the studies appear to be potentially eligible for inclusion. The retrieved full text articles will be independently reviewed by two authors (BD and MC). Disagreements will be resolved by consensus with a third author (DT).

Data extraction and management

Two investigators will independently complete a standardized data extraction form on all eligible studies. The following data will be retrieved:

1. General information: title, journal, year, published/unpublished;

2. Study information: design, sample size calculation (a priori and post hoc), and methods used to ensure adequacy of randomization, concealment of allocation and blinding;

3. Intervention: type, dose, formulation and frequency of administration of SCFAs and the control group, including explicit description of any add‐on therapies;

4. Eligibility: inclusion/ exclusion criteria, total number screened and randomized;

5. Baseline characteristics (in each group): age, sex, race, disease severity (and the method used for its evaluation), and concurrent medications;

6. Follow‐up: duration of follow‐up period, assessment of compliance, withdrawals and loss to follow‐up;

7. Outcome: proportion of patients achieving remission and clinical improvement at each evaluated time period (and the method used for its evaluation), time to remission, time to first relapse, change in quality of life measures, and explicit longitudinal description of adverse events; and

8. Scoring for the Cochrane risk of bias tool for assessing quality of eligible studies. 

Assessment of risk of bias in included studies

Included studies will be assessed for quality using the Cochrane risk of bias tool (Higgins 2008). This assessment tool addresses six main domains (sequence generation, allocation concealment, blinding, incomplete outcome data, selective outcome reporting and ‘other issues’). The risk of bias for each domain is assessed by answering pre‐specified questions about study characteristics (e.g. is the allocation sequence adequately generated?), in which ‘Yes’ indicates low risk of bias, ‘No’ indicates high risk of bias, and ‘Unclear’ indicates unclear or unknown risk of bias. Two authors (DB, MC) will independently score each study. Studies will be classified as high risk for bias, low risk for bias and unclear risk; disagreements will be resolved by consensus with a third author (DT).

The primary analysis will exclude all studies of low quality. Sensitivity analyses based on study quality are planned governed by the number of included trials.

Measures of treatment effect

Data for each study will be transformed in to 2X2 tables. Dichotomous outcomes (e.g. the proportions of patients entering remission or achieving clinical improvement) will be assessed using the relative risk and 95% confidence interval. Continuous outcomes (e.g. QoL) will be assessed using the weighted mean differences and 95% confidence interval. Data will be analyzed according to the intention to treat (ITT) principle.

Dealing with missing data

Every effort will be made to contact authors to obtain missing information. Sensitivity analyses will be performed while considering missing data in the interventional arms as failures and in the control arms as responders. Finally, studies with significant amount of missing data will be excluded in a sensitivity analysis. 

Assessment of heterogeneity

First, studies will be independently reviewed for any clinical and methodological heterogeneity and only similar studies will be combined. Possible reasons for heterogeneity will be discussed. The decision of whether to pool studies will be aided by calculating the I² statistics, interpreted as 25% (low heterogeneity), 50% (moderate) and 75% (high). Data will not be pooled if there is a large amount of unexplained heterogeneity (e.g. I²> 75%). The Cochran chi‐square test for homogeneity (Q test) will be calculated (a P value of 0.10 will be considered statistically significant).

Assessment of reporting biases

Funnel plots will be utilized to assess publication bias, if a sufficient number of studies are available.

Data synthesis

Data will be pooled for analysis if the patients, treatments and outcomes are sufficiently similar. Random or fixed effects models will be used to pool studies depending on clinical and statistical heterogeneity. Weighting will be assigned according to the Mantel‐Haenszel method.  The pooled absolute risk reduction and number needed to treat will be calculated where appropriate.

Subgroup analysis and investigation of heterogeneity

Subgroup analyses are planned (governed by the number of available studies) according to the different SCFA types, route, doses and frequency of administration, co‐interventions, duration of treatment, disease activity at baseline, and between children and adults. In addition, subgroup analyses are planned for the different outcome categories (i.e.  clinical, endoscopic or histological remission).

Sensitivity analysis

In order to assess the robustness of the results, sensitivity analyses will be performed excluding poor quality studies, and those published as abstracts only. Various extreme values will be used to explore the effect of missing data. For example, to assess attrition bias in cases of high drop‐out rate (>15%), those dropped from the intervention arm will be considered as failures and those from the comparison arm as success.

What's new

Date	Event	Description
20 June 2019	Amended	This protocol is being withdrawn

Declarations of interest

None known

Notes

This protocol is out of date and is being withdrawn

Withdrawn from publication for reasons stated in the review