Antibiotics for induction and maintenance of remission in Crohn's disease
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
To determine whether antibiotic therapy is safe and effective for induction or maintenance of remission in CD.
Background
Description of the condition
Crohn's disease (CD) is an inflammatory condition of the gastrointestinal tract that most commonly affects the ileum and the colon. Characteristic histologic features of the disease include transmural inflammation and mucosal ulceration. The exact etiology of CD is unclear, but environmental factors are important contributors (Elson 2005; Scribano 2013).
In animal models, interactions between the mucosal immune system and commensal bacteria contribute to the inflammatory changes seen in CD (Elson 1995; Elson 2005; Rath 1999). Subsequent human studies have demonstrated that patients with CD have higher concentrations of intestinal and colonic bacteria compared to healthy controls (Scribano 2013). Patients with CD may also have impaired barrier function that facilitates translocation of microbes into the mucosa (Marks 2006).
Pathogenic bacterial strains, including Escherichia coli (Mylonaki 2005), have been isolated in the mucosal and mesenteric lymph nodes of these patients (Ambrose 1984). Furthermore, there is a change in the microbial composition with fewer species overall and a relative overrepresentation of Enterobacteriaceae,Proteobacteria,Actinobacteria (Sartor 2008), and Bacteroides (Barnich 2007). This lends support to the theory that the pathologic immune responses in CD are in response to an inciting pathogen. This response is maintained through continued bacterial exposure, defects in the mucosal barrier, homology between the mucosa and bacteria and reduced tolerance in the mucosal immune system (Sartor 2008).
Description of the intervention
As a result of the suspected link between increased intestinal bacterial concentrations and chronic inflammation, antibiotics have been considered for the treatment of CD (Swidsinski 2002). Studies have suggested Escherichia coli as specific bacterial targets, among others (Mylonaki 2005; Sartor 2008).
How the intervention might work
Several antibiotics have been suspected to be of use in managing CD. By reducing the bacterial load in the intestinal mucosa, the inflammatory response to the pathogenic bacteria should be reduced (Scribano 2013; Swidsinski 2002). Furthermore, antibiotics also act to limit bacterial translocation and reduce the concentration of adherent bacteria to the lumen and mucosa (Scribano 2013). In patients who have high levels of Escherichia in their microbiome, treatment with mesalamine showed a decrease in intestinal inflammation. This further suggests the crucial role the gut microbiome may have in IBD pathophysiology and the potential use for antimicrobial agents (Kostic 2014).
Why it is important to do this review
Cumulatively these data have raised the possibility that alteration of the mucosal flora may have a therapeutic role in CD by inhibiting the stimulus for pathogenic immune responses (Ott 2004; Swidsinski 2002). Given these observations, it is sensible to postulate that antibiotic therapy might be effective for either induction or maintenance of remission in CD. However, several problems exist with this approach. First, use of broad‐spectrum antibiotics is a very blunt instrument and theoretically could actually aggravate the aforementioned 'dysbiosis'. Second, the resident flora is determined by both genetic and dietary factors that are difficult or impossible to modify on a chronic basis. Therefore, treatment, if effective, might have to be continued indefinitely. Finally broad‐spectrum antibiotic therapy is associated with important adverse effects, notably an increased risk of Clostridium difficile infection. For these reasons evidence from high quality randomized controlled trials (RCTs) is necessary before antibiotics are accepted as effective and safe for the treatment of CD.
Moreover, antibiotics have been used for the treatment of CD but the safety and efficacy remain unclear and there are conflicting recommendations among physicians. There are no recommendations regarding the antibiotic of choice, dose, or duration. The most recent guidelines published by the World Gastroenterology Organization support the use of antibiotics in perianal disease, fistulizing disease, and bacterial overgrowth secondary to stricturing disease, despite limited supporting evidence. There is extensive evidence regarding antibiotic use in post‐operative CD management (Bernstein 2016).
Objectives
To determine whether antibiotic therapy is safe and effective for induction or maintenance of remission in CD.
Methods
Criteria for considering studies for this review
Types of studies
RCTs of adult patients (> 18 years of age) will be considered for inclusion. Induction of remission studies need to have a minimum duration of at least four weeks to be considered for inclusion. Maintenance of remission studies need to have a minimum duration of at least six months to be considered for inclusion.
Types of participants
Patients with active or quiescent CD (as defined by the original studies) will be considered for inclusion.
Types of interventions
Trials that compare oral antibiotic therapy to a placebo or an active comparator will be considered for inclusion.
Types of outcome measures
Primary outcomes
The primary outcome measure for induction of remission studies will be the proportion of patients who failed to achieve remission (as defined by the original studies). The primary outcome for maintenance of remission studies will be the proportion of patients who relapse (as defined by the included studies).
Secondary outcomes
Secondary outcomes include the proportion of patients:
1. Who fail to achieve clinical response (as defined by the original studies);
2. Who fail to achieve endoscopic response (as defined by the original studies );
3. Who fail to achieve endoscopic remission (as defined by the original studies);
4.Who fail to achieve histological response (as defined by the original studies);
5.Who fail to achieve histological remission (as defined by the original studies);
6. Who fail to achieve both clinical and endoscopic response (as defined by the original studies);
7. Who fail to achieve both clinical and endoscopic remission (as defined by the original studies);
8. With adverse events;
9. With serious adverse events;
10. Who withdrew from the study due to adverse events; and
11. Quality of life (as measured by a validated quality of life instrument).
Search methods for identification of studies
Electronic searches
We will search the following databases for relevant studies:
1. MEDLINE (Ovid, 1946 to present);
2. EMBASE (Ovid, 1984 to present);
3. CENTRAL; and
4. The Cochrane IBD Group Specialized Register.
The search strategies are listed in Appendix 1.
Searching other resources
We will also search the references listed in relevant trials and review articles for additional studies not identified in the search. Furthermore, conference proceedings from major meetings (Digestive Disease Week, the European Crohn's and Colitis Organization congress, and the United European Gastroenterology Week conference) from the last five years will be searched for studies published in abstract form only.
Data collection and analysis
Selection of studies
Two authors (CMT and CEP) will screen the search results independently for eligible studies based on the inclusion criteria as listed. Disagreements will be discussed until a consensus is reached. Any disagreements will be brought to a third author (JKM) for resolution.
Data extraction and management
Data will be extracted from included studies by two independent authors (CMT and CEP). Any disagreements over extracted data will be first discussed and then brought to a third author (JKM) for resolution.
Assessment of risk of bias in included studies
The methodological quality of included studies will be independently assessed by two authors (CMT and CEP) using the Cochrane risk of bias tool (Higgins 2011). We will assess several factors including sequence generation, allocation sequence concealment, blinding, incomplete outcome data, selective outcome reporting and other potential sources of bias. Studies will be judged to be at high, low or unclear risk of bias. Any disagreements regarding risk of bias will be first discussed and then brought to a third author (JKM) for resolution.
We will use the GRADE approach to determine the overall quality of evidence supporting both primary and selected secondary outcomes (Guyatt 2008, Schünemann 2011). Evidence from RCTs will be considered high quality evidence. However, the quality of the evidence can be downgraded when considering the following factors:
1. Risk of bias;
2. Indirect evidence;
3. Inconsistency (unexplained heterogeneity);
4. Imprecision; and
5. Publication bias.
Each outcome will be reviewed to determine the overall quality of evidence supporting the outcome. The outcome will be classified as high quality (the estimate of effect is very unlikely to be changed despite further research); moderate quality (the estimate of effect is unlikely to be changed despite further research); low quality (the estimate of effect may be changed despite further research) or very low quality (the estimate of effect likely will be changed with further research).
Measures of treatment effect
Review Manager (RevMan 5.3.5) will be used to analyse the data on an intention‐to‐treat (ITT) basis. We will calculate the risk ratio (RR) and corresponding 95% confidence interval (95% CI) for dichotomous outcomes. We will calculate the mean difference (MD) and corresponding 95% CI for continuous outcomes.
Unit of analysis issues
To deal with repeated observations on participants, we will determine appropriate fixed intervals for follow‐up for each outcome. Cross‐over trials will be included if data is available for the first phase of the trial prior to cross‐over. To deal with events that may re‐occur (e.g. adverse events), we will report on the proportion of participants who experience at least one event. Seperate comparisons will be performed for studies that compare antibiotics to placebo and for studies that compare antibiotics to other active therapies. We will also perform separate comparisons for each type of antibiotic. If we encounter multiple treatment groups (e.g. different dose groups of antibiotics), we will divide the placebo group across the treatment groups or we may combine groups to create a single pair‐wise comparison as appropriate. We consider it to be unlikely that we will encounter any cluster‐randomized trials or studies where participants will receive multiple treatment attempts.
Dealing with missing data
We will use an intention‐to‐treat analysis for dichotomous outcomes whereby patients with missing treatment outcomes will be presumed to be treatment failures. We will perform sensitivity analyses to assess the impact of this assumption on the effect estimate.
Assessment of heterogeneity
We will assess heterogeneity using the Chi2 test (a P value of 0.10 will be considered statistically significant) and the I2 statistic. For the I2 statistic, 75% will indicate high heterogeneity among study data, 50% will indicate moderate heterogeneity and 25% will indicate low heterogeneity (Higgins 2003). Sensitivity analysis will be conducted to explore possible explanations for heterogeneity.
Assessment of reporting biases
We will initially compare outcomes listed in the protocol to those reported in the published manuscript. If we do not have access to the protocol, we will use the outcomes listed in the methods sections of the published manuscript compared to what is reported n the results section. If any pooled analyses include 10 or more studies, we will investigate potential publication bias using funnel plots (Egger 1997).
Data synthesis
We will combine data for meta‐analysis from individual trials when the interventions, patient groups and outcomes are similar as deemed by author consensus. We will calculate the pooled RR and corresponding 95% CI for dichotomous outcomes. We will calculate the pooled MD and corresponding 95% CI for continuous outcomes. We will calculate the standardized mean difference (SMD) and 95% CI when different scales are used to measure the same outcome. A fixed‐effect model will be used to pool data unless significant heterogeneity exists between the studies. A random‐effects model will be used if heterogeneity exits (I2 50 to 75%). We will not pool data for meta‐analysis if a high degree of heterogeneity (I2 ≥ 75%) is found.
Subgroup analysis and investigation of heterogeneity
Planned subgroup analysis (data allowing) will include:
a) Patient baseline characteristics (i.e. sex, age, weight, disease duration, disease severity, time since diagnosis, concomitant medication, objective markers of inflammation such as C‐reactive protein, and previous exposure to anti‐tumor necrosis factor‐alpha therapy); and
b) Different antibiotic doses .
Sensitivity analysis
We will use sensitivity analysis to assess the impact of random‐effects and fixed‐effect modelling, risk of bias, type of report (full manuscript, abstract or unpublished data) and loss to follow‐up on the pooled effect estimate.
