Antibiotics for induction and maintenance of remission in Crohn's disease

Cassandra M Townsend; Claire E Parker; John K MacDonald; Tran M Nguyen; Vipul Jairath; Brian G Feagan; Reena Khanna

doi:10.1002/14651858.CD012730.pub2

. 2019 Feb 7;2019(2):CD012730. doi: 10.1002/14651858.CD012730.pub2

Antibiotics for induction and maintenance of remission in Crohn's disease

Cassandra M Townsend ^1,^✉, Claire E Parker ², John K MacDonald ^1,³, Tran M Nguyen ³, Vipul Jairath ^1,⁴, Brian G Feagan ^1,^3,⁴, Reena Khanna ¹

Editor: Cochrane IBD Group

PMCID: PMC6366891 PMID: 30731030

Abstract

Background

Several antibiotics have been evaluated in Crohn's disease (CD), however randomised controlled trials (RCTs) have produced conflicting results.

Objectives

To assess the efficacy and safety of antibiotics for induction and maintenance of remission in CD.

Search methods

We searched MEDLINE, Embase, CENTRAL, the Cochrane IBD Group Specialized Register and Clinicaltrials.gov database from inception to 28 February 2018. We also searched reference lists and conference proceedings.

Selection criteria

RCTs comparing antibiotics to placebo or an active comparator in adult (> 15 years) CD patients were considered for inclusion.

Data collection and analysis

Two authors screened search results and extracted data. Bias was evaluated using the Cochrane risk of bias tool. The primary outcomes were failure to achieve clinical remission and relapse. Secondary outcomes included clinical response, endoscopic response, endoscopic remission, endoscopic relapse, histologic response, histologic remission, adverse events (AEs), serious AEs, withdrawal due to AEs and quality of life. Remission is commonly defined as a Crohn's disease activity index (CDAI) of < 150. Clinical response is commonly defined as a decrease in CDAI from baseline of 70 or 100 points. Relapse is defined as a CDAI > 150. For studies that enrolled participants with fistulizing CD, response was defined as a 50% reduction in draining fistulas. Remission was defined as complete closure of fistulas. We calculated the risk ratio (RR) and corresponding 95% confidence interval (95% CI) for dichotomous outcomes. We calculated the mean difference (MD) and corresponding 95% CI for continuous outcomes. GRADE was used to assess the certainty of the evidence.

Main results

Thirteen RCTs (N = 1303 participants) were eligible. Two trials were rated as high risk of bias (no blinding). Seven trials were rated as unclear risk of bias and four trials were rated as low risk of bias. Comparisons included ciprofloxacin (500 mg twice daily) versus placebo, rifaximin (800 to 2400 mg daily) versus placebo, metronidazole (400 mg to 500 mg twice daily) versus placebo, clarithromycin (1 g/day) versus placebo, cotrimoxazole (960 mg twice daily) versus placebo, ciprofloxacin (500 mg twice daily) and metronidazole (250 mg four time daily) versus methylprednisolone (0.7 to 1 mg/kg daily), ciprofloxacin (500 mg daily), metronidazole (500 mg daily) and budesonide (9 mg daily) versus placebo with budesonide (9 mg daily), ciprofloxacin (500 mg twice daily) versus mesalazine (2 g twice daily), ciprofloxacin (500 mg twice daily) with adalimumab versus placebo with adalimumab, ciprofloxacin (500 mg twice daily) with infliximab versus placebo with infliximab, clarithromycin (750 mg daily) and antimycobacterial versus placebo, and metronidazole (400 mg twice daily) and cotrimoxazole (960 mg twice daily) versus placebo. We pooled all antibiotics as a class versus placebo and antibiotics with anti‐tumour necrosis factor (anti‐TNF) versus placebo with anti‐TNF.

The effect of individual antibiotics on CD was generally uncertain due to imprecision. When we pooled antibiotics as a class, 55% (289/524) of antibiotic participants failed to achieve remission at 6 to 10 weeks compared with 64% (149/231) of placebo participants (RR 0.86, 95% CI 0.76 to 0.98; 7 studies; high certainty evidence). At 10 to 14 weeks, 41% (174/428) of antibiotic participants failed to achieve a clinical response compared to 49% (93/189) of placebo participants (RR 0.77, 95% CI 0.64 to 0.93; 5 studies; moderate certainty evidence). The effect of antibiotics on relapse in uncertain. Forty‐five per cent (37/83) of antibiotic participants relapsed at 52 weeks compared to 57% (41/72) of placebo participants (RR 0.87, 95% CI 0.52 to 1.47; 2 studies; low certainty evidence). Relapse of endoscopic remission was not reported in the included studies. Antibiotics do not appear to increase the risk of AEs. Thirty‐eight per cent (214/568) of antibiotic participants had at least one adverse event compared to 45% (128/284) of placebo participants (RR 0.87, 95% CI 0.75 to 1.02; 9 studies; high certainty evidence). The effect of antibiotics on serious AEs and withdrawal due to AEs was uncertain. Two per cent (6/377) of antibiotic participants had at least one adverse event compared to 0.7% (1/143) of placebo participants (RR 1.70, 95% CI 0.29 to 10.01; 3 studies; low certainty evidence). Nine per cent (53/569) of antibiotic participants withdrew due to AEs compared to 12% (36/289) of placebo participants (RR 0.86, 95% CI 0.57 to 1.29; 9 studies; low certainty evidence) is uncertain. Common adverse events in the studies included gastrointestinal upset, upper respiratory tract infection, abscess formation and headache, change in taste and paraesthesia

When we pooled antibiotics used with anti‐TNF, 21% (10/48) of patients on combination therapy failed to achieve a clinical response(50% closure of fistulas) or remission (closure of fistulas) at week 12 compared with 36% (19/52) of placebo and anti‐TNF participants (RR 0.57, 95% CI 0.29 to 1.10; 2 studies; low certainty evidence). These studies did not assess the effect of antibiotics and anti‐TNF on clinical or endoscopic relapse. Seventy‐seven per cent (37/48) of antibiotics and anti‐TNF participants had an AE compared to 83% (43/52) of anti‐TNF and placebo participants (RR 0.93, 95% CI 0.76 to 1.12; 2 studies, moderate certainty evidence). The effect of antibiotics and anti‐TNF on withdrawal due to AEs is uncertain. Six per cent (3/48) of antibiotics and anti‐TNF participants withdrew due to an AE compared to 8% (4/52) of anti‐TNF and placebo participants (RR 0.82, 95% CI 0.19 to 3.45; 2 studies, low certainty evidence). Common adverse events included nausea, vomiting, upper respiratory tract infections, change in taste, fatigue and headache

Authors' conclusions

Moderate to high quality evidence suggests that any benefit provided by antibiotics in active CD is likely to be modest and may not be clinically meaningful. High quality evidence suggests that there is no increased risk of adverse events with antibiotics compared to placebo. The effect of antibiotics on the risk of serious adverse events is uncertain. The effect of antibiotics on maintenance of remission in CD is uncertain. Thus, no firm conclusions regarding the efficacy and safety of antibiotics for maintenance of remission in CD can be drawn. More research is needed to determine the efficacy and safety of antibiotics as therapy in CD

Plain language summary

Antibiotics for the treatment of Crohn's disease

What is Crohn's disease?

Crohn's disease (CD) is an inflammatory disorder that can affect any segment of the gastrointestinal tract from the mouth to the anus. Common symptoms of CD include fever, diarrhea, abdominal pain and weight loss. CD is characterized by periods of relapse when people experience symptoms and periods of remission when the symptoms stop.

What are antibiotics?

Antibiotics are medications used to treat bacterial infections. Antibiotics are designed to target specific bacterial populations and have different mechanisms of action to stop a bacterial population from growing or eradicate the bacteria.

What is the purpose this study?

Antibiotics are commonly used for managing patients with CD because the inflammatory process in the bowel was believed to be triggered by a specific bacterial pathogen. Elimination of this bacterial target would allow the inflammatory process to resolve. However, current clinical guidelines do not recommend use of antibiotic agents to induce or maintain clinical remission in patients with CD because there is no definitive evidence to suggest a benefit to using antibiotics in this way.

How was this study performed?

A systematic review of current literature was performed to determine whether antibiotic therapy is effective to induce or maintain remission in CD. An electronic search of several databases was performed and studies that met our inclusion criteria were selected for further evaluation. Statistical analyses were performed to determine which specific antibiotics had an overall benefit.

What were the results?

Several antibiotics, including ciprofloxacin, metronidazole, clarithromycin, rifaximin and cotrimoxazole, have been studied in CD. Most of the included studies were small in size. When we pooled antibiotics as a class, these drugs provided a modest benefit over placebo (i.e. a fake drug such as a sugar pill) for induction of remission and improvement of CD symptoms. For example, remission rates were 45% (253/542) in participants who received antibiotics compared to 36% (82/231) in participants who received placebo. We rated the quality of evidence supporting this outcome as high. Few studies assessed the use of antibiotics for maintenance of remission in CD. The impact of antibiotics on preventing relapse in CD is uncertain. Antibiotics do not appear to increase the risk of side effects when compared to placebo. Common side effects reported in the studies included gastrointestinal upset, upper respiratory tract infection, abscess formation, headache, change in taste and paraesthesia (pins and needles in the extremities). Serious side effects were not well reported in the studies and the impact of antibiotics on the risk of serious side effects is uncertain.

Conclusions

Moderate to high quality evidence suggests that any benefit provided by antibiotics in active CD is likely to be very modest. High quality evidence suggests that there is no increased risk of side effects with antibiotics compared to placebo. The effect of antibiotics on the risk of serious side effects is uncertain. The effect of antibiotics on preventing relapse in CD is uncertain. Thus, no firm conclusions regarding the benefits and harms of antibiotics for maintenance of remission in CD can be drawn. More research is needed to determine the harms and benefits of antibiotic therapy in CD.

Summary of findings

Background

Description of the condition

Crohn's disease (CD) is an inflammatory disorder 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, however both genetic and environmental factors are important contributors (Elson 2005; Scribano 2013). In this regard, the human microbiome is considered to be a key environmental risk factor.

In animal models, interactions between the mucosal immune system and commensal bacteria contribute to the observed pathological 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 (Scribano 2013), and higher populations of specific bacteria (Gevers 2014), compared to healthy controls. 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). These observations support the notion that the pathological response in CD is driven by an abnormal response to the host microbiome and that manipulation of the flora though antibiotic treatment might be a potential therapy (Sartor 2008).

Description of the intervention

Given the proposed 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 evaluated for the treatment of CD. Reduction of the bacterial load in the intestinal mucosa might reduce the pathological immune response in the intestinal mucosa (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). 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).

Why it is important to do this review

Given the extensive animal and human data that support the role of bacteria in the pathogenesis of CD, it is reasonable to postulate that antibiotic therapy might be effective for either induction or maintenance of remission in CD. However, several potential problems exist with this approach. First, use of broad‐spectrum antibiotics is a very blunt strategy that may aggravate the aforementioned dysbiosis. Second, the resident flora are determined by both genetic and dietary factors that may be 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.

No current recommendations exist regarding the antibiotic of choice, dose, or duration for treatment of CD. The most recent guidelines published by the World Gastroenterology Organisation support the use of antibiotics in perianal disease, fistulizing disease, and bacterial overgrowth secondary to stricturing disease, despite limited supporting evidence (Bernstein 2016). There is 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 (> 15 years of age) were considered for inclusion. Induction of remission studies needed to have a minimum duration of at least four weeks to be considered for inclusion. Maintenance of remission studies needed 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) were considered for inclusion.

Types of interventions

Trials that compare oral antibiotic therapy to a placebo or an active comparator were considered for inclusion.

Types of outcome measures

Primary outcomes

The primary outcome measure for induction of remission studies was the proportion of patients who failed to achieve remission, as defined by the original studies. The primary outcome for maintenance of remission studies was the proportion of patients who relapsed, as defined by the included studies.

Secondary outcomes

Secondary efficacy outcomes, as defined by the original studies, were the proportion of patients:

1. Who failed to achieve clinical response (as defined by the original studies);

2. Who failed to achieve endoscopic response (as defined by the original studies);

3. Who failed to achieve endoscopic remission (as defined by the original studies);

4. Who failed to achieve histological response (as defined by the original studies);

5. Who failed to achieve histological remission (as defined by the original studies);

6. Who had an endoscopic relapse (as defined by the original studies);

7. Who failed to achieve both clinical and endoscopic response (as defined by the original studies);

8. Who failed to achieve both clinical and endoscopic remission (as defined by the original studies); and

9. Health‐related quality of life (as measured by a validated quality of life instrument).     Safety outcomes were the proportion of patients:

10. With any adverse event (AE);

11. With serious adverse events (SAE); and

12. Who withdrew from the study due to adverse events.

Search methods for identification of studies

Electronic searches

We searched 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.

5. Clinicaltrials.gov

The search strategies are listed in Appendix 1.

Searching other resources

We also searched the references listed in relevant studies and review articles for additional citations not identified in the search. Furthermore, conference proceedings from major meetings (Digestive Disease Week, the European Crohn's and Colitis Organisation congress, and the United European Gastroenterology Week conference) from the last five years were searched for studies published in abstract form only.

Data collection and analysis

Selection of studies

Two authors (CMT and CEP) screened the search results independently for eligible studies based on the inclusion criteria as listed. Disagreements were discussed until a consensus is reached. Any disagreements were brought to a third author (JKM) for resolution.

Data extraction and management

Data were extracted from included studies by two independent authors (CMT and CEP). Any disagreements over extracted data were first discussed and then brought to a third author (JKM) for resolution if deemed necessary.

Assessment of risk of bias in included studies

The methodological quality of included studies was independently assessed by two authors (CMT and CEP) using the Cochrane risk of bias tool (Higgins 2011). We assessed several factors including sequence generation, allocation sequence concealment, blinding, incomplete outcome data, selective outcome reporting and other potential sources of bias. Studies were judged to be at high, low or unclear risk of bias. Any disagreements regarding risk of bias were first discussed and then brought to a third author (JKM) for resolution.

We used the GRADE approach to determine the overall certainty of evidence supporting both primary and selected secondary outcomes (Guyatt 2008; Schünemann 2011). For the 'Summary of findings' tables, we included the following outcomes: failure to achieve clinical remission (at study endpoint), failure to maintain clinical remission (or relapse at study endpoint), failure to achieve clinical response (at study endpoint), failure to maintain endoscopic remission (or endoscopic relapse at study endpoint), adverse events, adverse events, serious adverse events and study withdrawal due to adverse events. Evidence from RCTs was considered high certainty. However, the certainty of the evidence could have been downgraded after considering the following factors:

1. Risk of bias;

2. Indirect evidence;

3. Inconsistency (unexplained heterogeneity);

4. Imprecision; and

5. Publication bias.

Each outcome was reviewed to determine the overall certainty of evidence supporting the outcome. The outcome was classified as high certainty (the estimate of effect is very unlikely to be changed despite further research); moderate certainty (the estimate of effect is unlikely to be changed despite further research); low certainty (the estimate of effect may be changed despite further research) or very low certainty (the estimate of effect likely will be changed with further research).

Measures of treatment effect

Review Manager (RevMan 5.3.5) was used to analyse the data on an intention‐to‐treat (ITT) basis. We calculated the risk ratio (RR) and corresponding 95% confidence interval (95% CI) for dichotomous outcomes and the mean difference (MD) and corresponding 95% CI for continuous outcomes.

Unit of analysis issues

To deal with repeated observations on participants, we determined appropriate fixed intervals for follow‐up for each outcome. Cross‐over trials were included if data was 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 reported on the proportion of participants who experience at least one event. Separate comparisons were performed for studies that compared antibiotics to placebo and for studies that compared antibiotics to other active therapies. We also performed separate comparisons for each type of antibiotic. If we encountered multiple treatment groups (e.g. different dose groups of antibiotics), we divided the placebo group across the treatment groups or we combined groups to create a single pair‐wise comparison as appropriate.

Dealing with missing data

An intention‐to‐treat analysis was used for dichotomous outcomes whereby patients with missing treatment outcomes were assumed to be treatment failures. Sensitivity analyses were performed to assess the impact of this assumption on the effect estimate.

Assessment of heterogeneity

Heterogeneity was assessed using the Chi² test (a P value of 0.10 was considered statistically significant) and the I² statistic. We considered an I² statistic> 75% to indicate high heterogeneity among study data, > 50% indicated moderate heterogeneity and > 25% will indicated low heterogeneity (Higgins 2003). Sensitivity analysis were conducted to explore possible explanations for heterogeneity.

Assessment of reporting biases

We initially compared outcomes listed in the protocol to those reported in the published manuscript. If we did not have access to the protocol, we used the outcomes listed in the methods sections of the published manuscript compared to what was reported in the results section. If any pooled analyses included 10 or more studies, we investigated potential publication bias using funnel plots (Egger 1997).

Data synthesis

Data for meta‐analysis from individual trials were combined when the interventions, patient groups and outcomes were similar, as deemed by author consensus. We calculated the pooled RR and corresponding 95% CI for dichotomous outcomes and the pooled MD and corresponding 95% CI for continuous outcomes. The standardized mean difference (SMD) and 95% CI was calculated when different scales were used to measure the same outcome. A fixed‐effect model was used to pool data unless significant heterogeneity existed between the studies. A random‐effects model was used if heterogeneity existed (I² = 50 to 75%). We did not pool data for meta‐analysis if a high degree of heterogeneity (I² ≥ 75%) was found.

Subgroup analysis and investigation of heterogeneity

Planned subgroup analysis (data allowing) included:

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‐tumour necrosis factor‐alpha therapy); and

b) Different antibiotic doses.

Sensitivity analysis

We planned to 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.

Results

Description of studies

Results of the search

The literature search conducted on 28 February 2018 retrieved 2334 records for consideration. We removed all duplicate records, which left 1803 records for screening. Two authors (CMT and CEP) reviewed the titles and abstracts independently and in duplicate. Forty‐eight articles were selected for full text review (see Figure 1). Thirty reports of 25 studies were excluded with reasons (See Characteristics of excluded studies). Seventeen reports of 13 trials met the inclusion criteria and were included in the review (See Characteristics of included studies). One ongoing study was identified (NCT02240108).

Included studies

Of the 13 eligible RCTs identified (N = 1303), five different antibiotics (ciprofloxacin, metronidazole, clarithromycin, rifaximin and cotrimoxazole) were evaluated. Eleven of these trials were placebo‐controlled (Ambrose 1985; Arnold 2002; Dewint 2014; Lieper 2008; Prantera 2006; Prantera 2012; Selby 2007; Steinhart 2002; Sutherland 1991; Thia 2009; West 2004), and two were active comparator trials (Colombel 1999; Prantera 1996). Two of the placebo‐controlled trials also included active comparator arms (Ambrose 1985; Thia 2009). Patients were adults with active CD at the time of randomisation. The majority of the included studies defined an adult as 18 years of age or older, however, Steinhart 2002, Ambrose 1985 and Thia 2009 included patients 14, 15 and 16 years of age or older, respectively.

Four placebo‐controlled RCTs (Arnold 2002; Dewint 2014; Thia 2009; West 2004) evaluated ciprofloxacin. One active comparator trial randomised patients to ciprofloxacin or oral mesalamine (Colombel 1999). In two studies, ciprofloxacin was administered in conjunction with an anti‐TNF agent (Dewint 2014; West 2004). In the Dewint 2014 study, all patients were treated with self‐administered adalimumab at an induction dose of 160 mg at day 0 and 80 mg at week 2, followed by maintenance of 40 mg every 4 weeks until week 24. In the West 2004 study, all participants received infliximab at a dose of 5 mg/kg at weeks 6, 8 and 12. Rifaximin was evaluated in two placebo‐controlled RCTs (Prantera 2006; Prantera 2012). Metronidazole was studied in three induction trials (Ambrose 1985; Steinhart 2002; Thia 2009) and one maintenance trial (Sutherland 1991). Two studies evaluated metronidazole in combination with other therapeutic agents. In Prantera 1996, patients were assigned to metronidazole combined with ciprofloxacin or methylprednisolone, while patients enrolled in Steinhart 2002 received metronidazole combined with ciprofloxacin or placebo. All participants in the Steinhart 2002 study received budesonide (9 mg/day). One study compared a combination of cotrimoxazole and metronidazole with placebo (Ambrose 1985). Two trials compared clarithromycin to placebo (Lieper 2008; Selby 2007). In Lieper 2008 patients were randomised to placebo or clarithromycin and followed for three months. Selby 2007 assigned patients to clarithromycin, oral rifabutin, oral clofazimine or placebo, in addition to a tapering course of prednisolone.

Excluded studies

Twenty‐five studies were excluded with reasons after the full text review was performed. In ten studies, data on outcomes of interest were not available in the manuscript, (Allan 1997; Biancone 1998; Goodgame 2001; Gui 1997; Hartley‐Asp 1981; Jigaranu 2014; Laudage 1983; Lee 2018; Mitelman 1982; Turunen 1995) Two of these studies were cross‐over trials that did not report on outcomes pre‐crossover (Blichfeldt 1978; Ursing 1982). Nine trials were not RCTs (Bernstein 1992; Gilat 1982; Jaworski 2016; Koretz 1997; Leiper 2000; Melmed 2009; Ronge 2007; Steele 2009; To 1995). Three trials was terminated and data were not available (Koch 2007; Rogler 2014; Steinhart 2008). One study evaluated rectal therapy, which was beyond the scope of this review (Maeda 2010).

Risk of bias in included studies

The risk of bias for the included studies is summarized in Figure 2. Overall, most studies received low or unclear risk of bias ratings for the for seven domains.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Six of the included studies did not adequately describe the methods used to for random sequence generation and therefore received an unclear risk of bias assessment for this domain (Ambrose 1985; Colombel 1999; Prantera 2006; Selby 2007; Sutherland 1991; West 2004). The remaining seven studies were rated as low risk of bias for this item (Arnold 2002; Dewint 2014; Lieper 2008; Prantera 1996; Prantera 2012; Steinhart 2002; Thia 2009).

Eight of the included studies did not adequately describe the methods used to conceal allocation and therefore received an unclear risk of bias rating for this domain (Ambrose 1985; Colombel 1999; Prantera 1996; Prantera 2006; Selby 2007; Steinhart 2002; Sutherland 1991; West 2004) . The remaining five studies received a low risk of bias rating for this item (Arnold 2002; Dewint 2014; Lieper 2008; Prantera 2012; Thia 2009) .

Blinding

One study did not adequately describe whether participants and personnel were blinded, and therefore received an unclear risk of bias rating for this domain (Ambrose 1985). A total of 10 studies were rated as low risk of bias for blinding of participants and personnel (Arnold 2002; Dewint 2014; Lieper 2008; Prantera 2006; Prantera 2012; Selby 2007; Steinhart 2002; Sutherland 1991; Thia 2009; West 2004). Two studies were rated as high risk of bias for this domain (Colombel 1999; Prantera 2006). In Colombel 1999, participants and investigators were not blinded. In Prantera 1996, patients were blinded but some investigators were unblinded.

It was unclear whether outcome assessors were blinded in three studies (Ambrose 1985; Arnold 2002; Thia 2009). Eight of the included studies were rated as low risk of bias for blinded of outcome assessment (Dewint 2014; Lieper 2008; Prantera 2006; Prantera 2012; Selby 2007; Steinhart 2002; Sutherland 1991; West 2004), while two studies did not employ blinded outcome assessment and received high risk of bias ratings (Colombel 1999; Prantera 1996).

Incomplete outcome data

Seven of the included studies were rated as unclear risk of bias with regard to incomplete outcome data (Ambrose 1985; Arnold 2002; Lieper 2008; Prantera 1996; Prantera 2012; Sutherland 1991; Thia 2009). The remaining six studies were rated as low risk of bias (Colombel 1999; Dewint 2014; Prantera 2006; Selby 2007; Steinhart 2002; West 2004).

Summary of findings for the main comparison. Antibiotic compared to placebo for induction and maintenance of remission in Crohn's disease.

Antibiotic compared to placebo for induction and maintenance of remission in Crohn's disease
Patient or population: Participants with active CD  Setting: Outpatient  Intervention: Antibiotic  Comparison: Placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Outcomes	Risk with placebo	Risk with Antibiotic	Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Failure to enter clinical remission Follow‐up: 6‐10 weeks	645 per 1,000	555 per 1,000  (490 to 632)	RR 0.86  (0.76 to 0.98)	773  (7 RCTs)	⊕⊕⊕⊕  HIGH	Clinical remission was defined as CDAI ≤150 Antibiotics included Cotrimoxazole, Metronidazole, Ciprofloxacin, Clarithromycin, and Rifaximin
Failure to maintain clinical remission Follow‐up: 52 weeks	569 per 1,000	495 per 1,000  (296 to 837)	RR 0.87  (0.52 to 1.47)	155  (2 RCTs)	⊕⊕⊝⊝  LOW ^{1 2}	Clinical remission was defined as CDAI ≤150 Antibiotics included Cotrimoxazole and Clarithromycin
Failure to achieve clinical response Follow‐up: 10‐14 weeks	492 per 1,000	379 per 1,000  (315 to 458)	RR 0.77  (0.64 to 0.93)	617  (5 RCTs)	⊕⊕⊕⊝  MODERATE ³	Clinical response was defined as a reduction in CDAI score of 100 points and/or a 50% or greater reduction in perianal fistulas Antibiotics included Ciprofloxacin and Rifaximin
Failure to maintain endoscopic remission	Not reported					This outcome was not reported
Adverse events Follow‐up: 6‐52 weeks	451 per 1,000	392 per 1,000  (338 to 460)	RR 0.87  (0.75 to 1.02)	852  (9 RCTs)	⊕⊕⊕⊕  HIGH	Adverse events included gastrointestinal upset, upper respiratory tract infection, abscess formation, headache and paraesthesia Antibiotics included Cotrimoxazole, Metronidazole, Ciprofloxacin, Clarithromycin, and Rifaximin
Serious adverse events Follow‐up: 6‐52 weeks	7 per 1,000	12 per 1,000  (2 to 70)	RR 1.70  (0.29 to 10.01)	520  (3 RCTs)	⊕⊕⊝⊝  LOW⁴	Serious adverse events were not well described in the studies. Reported serious adverse events included one scrotal edema and one death Antibiotics included Rifaximin, Ciprofloxacin and Metronidazole
Withdrawal due to adverse events Follow‐up: 6‐52 weeks	125 per 1,000	107 per 1,000  (71 to 161)	RR 0.86  (0.57 to 1.29)	858  (9 RCTs)	⊕⊕⊝⊝  LOW ⁵	Adverse events leading to withdrawal included worsening CD, gastrointestinal symptoms,headache, abscess, rash, arthralgia, nausea, vomiting, arthropathy and infusion reaction Antibiotics included Cotrimoxazole, Metronidazole, Ciprofloxacin, Clarithromycin, and Rifaximin
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Antibiotic with anti‐TNF compared to placebo with anti‐TNF for induction of remission in Crohn's disease
Patient or population: Participants with active CD  Setting: Outpatient  Intervention: Antibiotic with anti‐TNF  Comparison: Placebo with anti‐TNF
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with placebo with anti‐TNF	Risk with Antibiotic with anti‐TNF
Failure to enter clinical remission	Not reported					This outcome was reported in one study. We decided to pool this study with the other anti‐TNF study below (failure to achieve clinical response or remission Antibiotics included Ciprofloxacin
Failure to maintain clinical remission	Not reported					This outcome was not reported
Failure to achieve clinical response or remission Follow‐up: 12 weeks	365 per 1,000	208 per 1,000  (106 to 402)	RR 0.57  (0.29 to 1.10)	100  (2 RCTs)	⊕⊕⊝⊝  LOW ¹	Clinical response was defined as a 50% reduction in perianal fistulas. Remission was defined as a closure of fistulas Antibiotics included Ciprofloxacin
Failure to maintain endoscopic remission	Not reported					This outcome was not reported
Adverse events Follow‐up: 12 weeks	827 per 1,000	769 per 1,000  (628 to 926)	RR 0.93  (0.76 to 1.12)	100  (2 RCTs)	⊕⊕⊕⊝  MODERATE ²	Adverse events included nausea, vomiting, upper respiratory tract infections, fatigue and headache Antibiotics included Ciprofloxacin
Serious adverse events	Not reported					This outcome was not reported
Withdrawal due to adverse events Follow‐up: 12 weeks	77 per 1,000	63 per 1,000  (15 to 265)	RR 0.82  (0.19 to 3.45)	100  (2 RCTs)	⊕⊕⊝⊝  LOW ³	Adverse events leading to withdrawal included gastrointestinal symptoms, transfusion reaction and herpes simplex virus infection Antibiotics included Ciprofloxacin
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Ciprofloxacin compared to placebo for induction of remission in Crohn's disease
Patient or population: Participants with active CD  Setting: Outpatient  Intervention: Ciprofloxacin  Comparison: Placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Outcomes	Risk with placebo	Risk with Ciprofloxacin	Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Failure to enter clinical remission Follow‐up:10‐12 weeks	741 per 1,000	489 per 1,000  (311 to 770)	RR 0.66  (0.42 to 1.04)	65  (2 RCTs)	⊕⊕⊝⊝  LOW ¹	Clinical remission was defined as CDAI ≤150
Failure to maintain clinical remission Follow‐up: 24 weeks	842 per 1,000	320 per 1,000 (185 to 573)	RR 0.38 (0.22 to 0.68)	47 (1 RCT)	⊕⊕⊝⊝  LOW ²	Clinical remission was defined as CDAI ≤150
Failure to have clinical response Follow‐up: 10 weeks	875 per 1,000	403 per 1,000 (175 to 893)	RR 0.46 (0.20 to 1.02)	18 (1 RCT)	⊕⊕⊝⊝  LOW ³	Clinical response was defined as at least a 50% reduction in baseline draining fistulas
Failure to maintain endoscopic remission	Not reported					This outcome was not reported
Adverse events Follow‐up: 10‐24 weeks	407 per 1,000	407 per 1,000  (232 to 717)	RR 1.00  (0.57 to 1.76)	65  (2 RCTs)	⊕⊕⊝⊝  LOW ⁴	Adverse events included upper respiratory tract infection, abscess, open fistula, arthralgias and unpleasant taste/sore mouth
Serious adverse events	0 per 1,000	0 per 1,000 (0 to 0)	not  estimable	18  (1 RCT)		No serious adverse events were observed
Withdrawal due to adverse events Follow‐up: 10‐24 weeks	148 per 1,000	50 per 1,000  (10 to 247)	RR 0.34  (0.07 to 1.67)	65  (2 RCTs)	⊕⊕⊝⊝  LOW ⁵	Withdrawals were due to worsening Crohn's disease
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Rifaximin compared to placebo for induction of remission in Crohn's disease
Patient or population: Participants with active CD  Setting: Outpatient  Intervention: Rifaximin  Comparison: Placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Outcomes	Risk with placebo	Risk with Rifaximin	Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Failure to enter clinical remission Follow‐up 12‐14 weeks	597 per 1,000	489 per 1,000 (412 to 585)	RR 0.82 (0.69 to 0.98)	489 (2 RCTs)	⊕⊕⊕⊝  MODERATE¹	Clinical remission was defined as CDAI ≤150
Failure to maintain clinical remission	Not reported					This outcome was not reported
Failure to have clinical response Follow‐up: 12‐14 weeks	519 per 1,000	426 per 1,000  (348 to 525)	RR 0.82  (0.67 to 1.01)	489  (2 RCTs)	⊕⊕⊕⊝  MODERATE²	Clinical response was defined as reduction of CDAI ≥ 70 points and reduction in CDAI score of 100 points
Failure to maintain endoscopic remission	Not reported					This outcome was not reported
Adverse events Follow‐up: 12‐14 weeks	473 per 1,000	392 per 1,000  (312 to 492)	RR 0.83  (0.66 to 1.04)	489  (2 RCTs)	⊕⊕⊕⊝  MODERATE ³	Adverse events included gastrointestinal disorders, infections, headache and ocular disorders
Serious adverse events Follow‐up: 12‐14 weeks	8 per 1,000	9 per 1,000  (2 to 35)	RR 1.11  (0.27 to 4.54)	489  (2 RCTs)	⊕⊕⊝⊝  LOW ⁴	The types of serious adverse events were not described by study authors
Withdrawal due to adverse events Follow‐up: 12 ‐14 weeks	62 per 1,000	78 per 1,000  (37 to 164)	RR 1.25  (0.59 to 2.64)	489  (2 RCTs)	⊕⊕⊝⊝  LOW ⁵	The adverse events leading to withdrawal were not described by study authors
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Metronidazole compared to placebo for induction of remission in Crohn's disease
Patient or population: Participants with active CD  Setting: Outpatient  Intervention: Metronidazole  Comparison: Placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Outcomes	Risk with placebo	Risk with Metronidazole	Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Failure to enter clinical remission Follow‐up: 6‐10 weeks	680 per 1,000	619 per 1,000  (422 to 904)	RR 0.91  (0.62 to 1.33)	50  (2 RCTs)	⊕⊕⊝⊝  LOW ¹	Clinical remission was defined as closure of all open actively draining fistulas at baseline
Failure to maintain clinical remission	Not reported					This outcome was not reported
Failure to have clinical response Follow‐up: 10 weeks	875 per 1,000	604 per 1,000 (341 to 1,000)	RR 0.69 (0.39 to 1.21)	18 (1 RCT)	⊕⊕⊝⊝  LOW ²	Clinical response was defined as at least a 50% reduction in baseline draining fistulas
Failure to maintain endoscopic remission	Not reported					This outcome was not reported
Adverse events Follow‐up: 6‐10 weeks	274 per 1,000	233 per 1,000  (88 to 633)	RR 0.85  (0.32 to 2.31)	149  (3 RCTs)	⊕⊝⊝⊝  VERY LOW ^{3 4}	Adverse events include gastrointestinal upset, abscess formation and arthropathy.paraesthesias and sore mouth.
Serious adverse events	0 per 1,000	0 per 1,000  (0 to 0)	not estimable	15  (1 RCT)		No serious adverse events were observed
Withdrawal due to adverse events Follow‐up: 6‐10 weeks	148 per 1,000	114 per 1,000  (53 to 248)	RR 0.77  (0.36 to 1.68)	149  (3 RCTs)	⊕⊕⊝⊝  LOW ⁵	Withdrawal due to adverse events was most often due to headache, gastrointestinal symptoms, abscess formation, rash and arthralgia
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Clarithromycin compared to placebo for induction of remission in Crohn's disease
Patient or population: Participants with active CD  Setting: Outpatient  Intervention: Clarithromycin  Comparison: Placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Outcomes	Risk with placebo	Risk with Clarithromycin	Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Failure to enter clinical remission Follow‐up:12 weeks	818 per 1,000	843 per 1,000  (638 to 1,000)	RR 1.03  (0.78 to 1.36)	41  (1 RCT)	⊕⊕⊝⊝  LOW ¹	Clinical remission was defined as CDAI ≤150
Failure to maintain clinical remission	Not reported					This outcome was not reported
Failure to have clinical response Follow‐up: 12 weeks	818 per 1,000	736 per 1,000 (532 to 1,000)	RR 0.90 (0.65 to 1.26)	41 (1 RCT)	⊕⊕⊝⊝  LOW ²	Clinical response was defined by CDAI reduction by ≥ 70 from baseline
Failure to maintain endoscopic remission	Not reported					This outcome was not reported
Adverse events Follow‐up: 12 weeks	45 per 1,000	210 per 1,000  (26 to 1,000)	RR 4.63  (0.57 to 37.96)	41  (1 RCT)	⊕⊕⊝⊝  LOW ³	Adverse events included gastrointestinal symptoms
Serious adverse events	Not reported					This outcome was not reported
Withdrawal due to adverse events Follow‐up: 12 weeks	500 per 1,000	370 per 1,000  (180 to 760)	RR 0.74  (0.36 to 1.52)	41  (1 RCT)	⊕⊕⊝⊝  LOW ⁴	Withdrawal due to adverse events was most often due to gastrointestinal symptoms
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Cotrimoxazole compared to placebo for induction of remission in Crohn's disease
Patient or population: Participants with active CD  Setting: Outpatient  Intervention: Cotrimoxazole  Comparison: Placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with placebo	Risk with Cotrimoxazole
Failure to enter clinical remission Follow‐up: week 12	588 per 1,000	688 per 1,000  (412 to 1,000)	RR 1.17  (0.70 to 1.96)	33  (1 RCT)	⊕⊕⊝⊝  LOW ¹	Clinical remission was defined as improvement in clinical assessment and laboratory indices
Failure to maintain clinical remission	Not reported					This outcome was not reported
Failure to have a clinical response	Not reported					This outcome was not reported
Failure to maintain endoscopic remission	Not reported					This outcome was not reported
Adverse events Follow‐up: 12 weeks	176 per 1,000	125 per 1,000  (25 to 653)	RR 0.71  (0.14 to 3.70)	33  (1 RCT)	⊕⊕⊝⊝  LOW ²	Adverse events included nausea, vomiting and arthropathy
Serious adverse events	Not reported					This outcome was not reported
Withdrawal due to adverse events Follow‐up: 12 weeks	59 per 1,000	125 per 1,000  (12 to 1,000)	RR 2.13  (0.21 to 21.22)	33  (1 RCT)	⊕⊕⊝⊝  LOW ³	Withdrawal due to adverse events was most often due to nausea, vomiting and arthropathy
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Ciprofloxacin and metronidazole compared to methylprednisone for induction and maintenance of remission in Crohn's disease
Patient or population: Participants with active CD  Setting: Outpatient  Intervention: Ciprofloxacin and metronidazole  Comparison: Methylprednisone
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Outcomes	Risk with methylprednisone	Risk with Ciprofloxacin and metronidazole	Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Failure to enter clinical remission Follow‐up: 12 weeks	368 per 1,000	545 per 1,000  (269 to 1,000)	RR 1.48  (0.73 to 2.99)	41  (1 RCT)	⊕⊝⊝⊝  VERY LOW^{1 2}	Clinical remission was defined as CDAI ≤150
Failure to maintain clinical remission Follow‐up: 52 weeks	684 per 1,000	773 per 1,000  (527 to 1,000)	RR 1.13  (0.77 to 1.65)	41  (1 RCT)	⊕⊕⊝⊝  VERY LOW^{2 3}	Clinical remission was defined as CDAI ≤150
Failure to have clinical response	Not reported					This outcome was not reported
Failure to maintain endoscopic remission	Not reported					This outcome was not reported
Adverse events Follow‐up: 12‐52 weeks	105 per 1,000	273 per 1,000  (62 to 1,000)	RR 2.59  (0.59 to 11.36)	41  (1 RCT)	⊕⊕⊝⊝  LOW ³	Adverse events include nausea, metallic taste, reflux symptoms, Cushingoid facies and acne
Serious adverse events	Not reported					This outcome was not reported
Withdrawal due to adverse events Follow‐up: 12‐52 weeks	105 per 1,000	273 per 1,000  (62 to 1,000)	RR 2.59  (0.59 to 11.36)	41  (1 RCT)	⊕⊕⊝⊝  LOW ³	Withdrawal due to adverse events was most often due to nausea, vomiting, reflux symptoms, hypertension, elevated amylase, acne and tremor
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Ciprofloxacin and metronidazole and budesonide compared to placebo with budesonide for induction of remission in Crohn's disease
Patient or population: Participants with active CD  Setting: Outpatient  Intervention: Ciprofloxacin and metronidazole  Comparison: Placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Outcomes	Risk with placebo	Risk with Ciprofloxacin and metronidazole	Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Failure to enter clinical remission Follow‐up: 8 weeks	632 per 1,000	683 per 1,000  (531 to 873)	RR 1.08  (0.84 to 1.38)	134  (1 RCT)	⊕⊕⊕⊝  MODERATE ¹	Clinical remission was defined as CDAI ≤150
Failure to maintain clinical remission	Not reported					This outcome was not reported
Failure to have clinical response	Not reported					This outcome was not reported
Failure to maintain endoscopic remission	Not reported					This outcome was not reported
Adverse events Follow‐up: 8 weeks	0 per 1,000	0 per 1,000  (0 to 0)	RR 27.81  (1.69 to 458.44)	134  (1 RCT)	⊕⊕⊝⊝  LOW ²	Adverse events included taste disturbance, dizziness, gastrointestinal upset and vaginitis
Serious adverse events Follow‐up: 8 weeks	29 per 1,000	46 per 1,000  (8 to 263)	RR 1.55  (0.27 to 8.95)	134  (1 RCT)	⊕⊕⊝⊝  LOW ³	The types of serious adverse events were not reported
Withdrawal due to adverse events Follow‐up: 8 weeks	0 per 1,000	0 per 1,000  (0 to 0)	RR 27.81  (1.69 to 458.44)	134  (1 RCT)	⊕⊕⊝⊝  LOW ²	Withdrawals due to adverse events were most often due to nausea, taste disturbance and rash
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Ciprofloxacin compared to mesalazine for induction of remission in Crohn's disease
Patient or population: Participants with active CD  Setting: Outpatient  Intervention: Ciprofloxacin  Comparison: Mesalazine
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with mesalazine	Risk with Ciprofloxacin
Failure to enter clinical remission Follow‐up: 6 weeks	455 per 1,000	445 per 1,000  (223 to 886)	RR 0.98  (0.49 to 1.95)	40  (1 RCT)	⊕⊝⊝⊝  VERY LOW ^{1 2}	Clinical remission was defined as CDAI ≤150
Failure to maintain clinical remission	Not reported					This outcome was not reported
Failure to have clinical response	Not reported					This outcome was not reported
Failure to maintain endoscopic remission	Not reported					This outcome was not reported
Adverse Events Follow‐up: 6 weeks	0 per 1,000	0 per 1,000  (0 to 0)	RR 3.63  (0.16 to 84.11)	40  (1 RCT)	⊕⊝⊝⊝  VERY LOW ^{1 3}	Adverse events included abdominal pain
Serious adverse events	Not reported					This outcome was not reported
Withdrawal due to adverse events Follow‐up: 6 weeks	0 per 1,000	0 per 1,000  (0 to 0)	RR 3.63  (0.16 to 84.11)	40  (1 RCT)	⊕⊝⊝⊝  VERY LOW ^{1 3}	The withdrawal due to an adverse event was due to abdominal pain
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Ciprofloxacin with adalimumab compared to placebo with adalimumab for induction and maintenance of remission in Crohn's disease
Patient or population: Participants with active CD  Setting: Outpatient  Intervention: Ciprofloxacin with adalimumab  Comparison: Placebo with adalimumab
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Outcomes	Risk with placebo with adalimumab	Risk with Ciprofloxacin with adalimumab	Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Failure to enter clinical remission Follow‐up: 12 weeks	359 per 1,000	244 per 1,000  (118 to 492)	RR 0.68  (0.33 to 1.37)	76  (1 RCT)	⊕⊕⊝⊝  LOW ¹	Clinical remission was defined as the closure of all fistulas All patients were treated with self‐administered adalimumab (patients were given induction dosing of 160 mg at day 0 and 80 mg at week 2, followed by maintenance of 40 mg every 4 weeks until week 24)
Failure to maintain clinical remission Follow‐up: 24 weeks	See comments					Although the authors reported on this outcome, we did not include it as participants only received ciprofloxacin up to week 12. All participants received maintenance adalimumab after week 12
Failure to have clinical response	Not reported					This outcome was not reported
Failure to maintain endoscopic remission	Not reported					This outcome was not reported
Adverse Events Follow‐up: 12‐24 weeks	872 per 1,000	837 per 1,000  (697 to 1,000)	RR 0.96  (0.80 to 1.16)	76  (1 RCT)	⊕⊕⊕⊝  MODERATE ³	Adverse events included respiratory tract infection, fatigue and headache
Serious adverse events Follow‐up: 12‐24 weeks	77 per 1,000	81 per 1,000  (18 to 377)	RR 1.05  (0.23 to 4.90)	76  (1 RCT)	⊕⊕⊝⊝  LOW ⁴	Serious adverse events included sagittal sinus thrombosis, severe disease flares, herpes simplex infection and parastomal herniation
Withdrawal due to adverse events Follow‐up: 12‐24 weeks	77 per 1,000	54 per 1,000  (9 to 305)	RR 0.70  (0.12 to 3.97)	76  (1 RCT)	⊕⊕⊝⊝  LOW ⁵	Specific adverse events causing withdrawal were not reported
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to enter clinical remission week 12 or 14	2	489	Risk Ratio (M‐H, Fixed, 95% CI)	0.82 [0.69, 0.98]
1.1 Dose 800 mg daily	2	178	Risk Ratio (M‐H, Fixed, 95% CI)	0.85 [0.65, 1.12]
1.2 Dose 1600 mg daily	2	174	Risk Ratio (M‐H, Fixed, 95% CI)	0.68 [0.50, 0.93]
1.3 Dose 2400 mg daily	1	137	Risk Ratio (M‐H, Fixed, 95% CI)	0.97 [0.69, 1.38]
2 Failure to have clinical response at week 12 or 14	2	489	Risk Ratio (M‐H, Fixed, 95% CI)	0.82 [0.67, 1.01]
2.1 Dose 800 mg daily	2	178	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.65, 1.28]
2.2 Dose 1600 mg daily	2	174	Risk Ratio (M‐H, Fixed, 95% CI)	0.63 [0.43, 0.91]
2.3 Dose 2400 mg daily	1	137	Risk Ratio (M‐H, Fixed, 95% CI)	0.97 [0.67, 1.40]
3 Adverse events	2	489	Risk Ratio (M‐H, Fixed, 95% CI)	0.83 [0.66, 1.04]
3.1 Dose 800 mg daily	2	178	Risk Ratio (M‐H, Fixed, 95% CI)	0.72 [0.49, 1.05]
3.2 Dose 1600 mg daily	2	174	Risk Ratio (M‐H, Fixed, 95% CI)	0.84 [0.58, 1.21]
3.3 Dose 2400 mg daily	1	137	Risk Ratio (M‐H, Fixed, 95% CI)	0.99 [0.64, 1.53]
4 Serious adverse events	2	489	Risk Ratio (M‐H, Fixed, 95% CI)	1.11 [0.27, 4.54]
4.1 Dose 800 mg daily	2	178	Risk Ratio (M‐H, Fixed, 95% CI)	1.64 [0.08, 33.26]
4.2 Dose 1600 mg daily	2	174	Risk Ratio (M‐H, Fixed, 95% CI)	1.31 [0.14, 12.08]
4.3 Dose 2400 mg daily	1	137	Risk Ratio (M‐H, Fixed, 95% CI)	0.66 [0.06, 7.05]
5 Withdrawal due to adverse events	2	489	Risk Ratio (M‐H, Fixed, 95% CI)	1.25 [0.59, 2.64]
5.1 Dose 800 mg daily	2	178	Risk Ratio (M‐H, Fixed, 95% CI)	0.56 [0.14, 2.16]
5.2 Dose 1600 mg daily	2	174	Risk Ratio (M‐H, Fixed, 95% CI)	1.07 [0.30, 3.83]
5.3 Dose 2400 mg daily	1	137	Risk Ratio (M‐H, Fixed, 95% CI)	2.64 [0.64, 10.90]

Ciprofloxacin with infliximab compared to placebo with infliximab for induction of remission in Crohn's disease
Patient or population: Participants with active CD  Setting: Outpatient  Intervention: Ciprofloxacin with infliximab  Comparison: Placebo with infliximab
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with placebo with infliximab	Risk with Ciprofloxacin with infliximab
Failure to enter clinical remission	Not reported					This outcome was not reported
Failure to maintain clinical remission	Not reported					This outcome was not reported
Failure to have clinical response Follow‐up: 12 weeks	385 per 1,000	92 per 1,000  (12 to 665)	RR 0.24  (0.03 to 1.73)	24  (1 RCT)	⊕⊕⊝⊝  LOW ¹	Clinical response was defined as 50% or greater reduction in draining fistulae confirmed by no drainage despite firm finger compression
Failure to maintain endoscopic remission	Not reported					This outcome was not reported
Adverse events Follow‐up: 12 weeks	692 per 1,000	547 per 1,000  (284 to 1,000)	RR 0.79  (0.41 to 1.51)	24  (1 RCT)	⊕⊕⊝⊝  LOW ²	Adverse events included nausea, rash and diarrhea and metallic taste
Serious adverse events	Not reported					This outcome was not reported
Withdrawal due to adverse events Follow‐up: 12 weeks	77 per 1,000	91 per 1,000  (6 to 1,000)	RR 1.18  (0.08 to 16.78)	24  (1 RCT)	⊕⊕⊝⊝  LOW ³	Adverse events leading to withdrawal included infusion reaction and disease exacerbation
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Clarithromycin and antimycobacterial compared to placebo for maintenance of remission in Crohn's disease
Patient or population: Participants with active CD  Setting: Outpatient  Intervention: Clarithromycin and antimycobacterial  Comparison: Placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Outcomes	Risk with placebo	Risk with Clarithromycin and antimycobacterial	Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Failure to enter clinical remission Follow‐up: 16 weeks	505 per 1,000	343 per 1,000  (247 to 474)	RR 0.68  (0.49 to 0.94)	213  (1 RCT)	⊕⊕⊕⊝  MODERATE ¹	Clinical remission was defined as CDAI ≤150
Failure to maintain clinical remission Follow‐up: 1 year	564 per 1,000	389 per 1,000  (265 to 569)	RR 0.69  (0.47 to 1.01)	122  (1 RCT)	⊕⊕⊕⊝  MODERATE ²	Clinical remission was defined as CDAI ≤150
Failure to have clinical response	Not reported					This outcome was not reported
Failure to maintain endoscopic remission Follow‐up: 3 years	800 per 1,000	864 per 1,000  (736 to 1,000)	RR 1.08  (0.92 to 1.27)	122  (1 RCT)	⊕⊕⊕⊝  MODERATE ³	The definition of endoscopic remission was not reported
Adverse events	Not reported					This outcome was not reported
Serious adverse events	Not reported					This outcome was not reported
Withdrawal of adverse events Follow‐up: 16 weeks‐3 years	73 per 1,000	75 per 1,000  (21 to 265)	RR 1.03  (0.29 to 3.64)	122  (1 RCT)	⊕⊕⊝⊝  LOW ⁴	Adverse events leading to withdrawal included abdominal distention, abnormal liver enzymes, vaginal candidiasis and myalgia
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Metronidazole and cotrimoxazole compared to placebo for induction of remission in Crohn's disease
Patient or population: Participants with active CD  Setting: Outpatient  Intervention: Metronidazole and cotrimoxazole  Comparison: Placebo
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Outcomes	Risk with placebo	Risk with Metronidazole and cotrimoxazole	Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
Failure to enter clinical remission Follow‐up: 6 weeks	588 per 1,000	618 per 1,000  (371 to 1,000)	RR 1.05  (0.63 to 1.77)	38  (1 RCT)	⊕⊕⊝⊝  LOW ¹	Clinical remission was based on clinical assessment and laboratory indices
Failure to maintain clinical remission	Not reported					This outcome was not reported
Failure to have clinical response	Not reported					This outcome was not reported
Failure to maintain endoscopic remission	Not reported					This outcome was not reported
Adverse events Follow‐up: 6 weeks	235 per 1,000	47 per 1,000  (5 to 388)	RR 0.20  (0.02 to 1.65)	38  (1 RCT)	⊕⊕⊝⊝  LOW ²	Adverse events included nausea, vomiting, esophagitis and erythema nodosum
Serious adverse events Follow‐up: 6 weeks	59 per 1,000	16 per 1,000  (1 to 371)	RR 0.27  (0.01 to 6.30)	38  (1 RCT)	⊕⊕⊝⊝  LOW ³	Serious adverse events included surgery abscess, nausea, vomitting and erythema nodosum
Withdrawal due to adverse events Follow‐up: 6 weeks	59 per 1,000	48 per 1,000  (3 to 706)	RR 0.81  (0.05 to 12.01)	38  (1 RCT)	⊕⊕⊝⊝  LOW ⁴	Adverse events leading to withdrawal included surgery abscess, nausea, vomiting and erythema nodosum
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio; OR: Odds ratio;
GRADE Working Group grades of evidence  High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to enter clinical remission at week 10 or 12	2	65	Risk Ratio (M‐H, Fixed, 95% CI)	0.61 [0.41, 0.92]
2 Failure to maintain clinical remission at week 24	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
3 Failure to have a clinical response at week 10	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
4 Adverse events	2	65	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.57, 1.76]
5 Serious adverse events	1	18	Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
6 Withdrawal due to adverse events	2	65	Risk Ratio (M‐H, Fixed, 95% CI)	0.34 [0.07, 1.67]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to enter clinical remission week 6 or 10	2	50	Risk Ratio (M‐H, Fixed, 95% CI)	0.91 [0.62, 1.33]
2 Failure to enter clinical remission at week 16	1	99	Risk Ratio (M‐H, Fixed, 95% CI)	1.03 [0.77, 1.36]
2.1 Dose 10 mg/kg	1	51	Risk Ratio (M‐H, Fixed, 95% CI)	0.95 [0.63, 1.45]
2.2 Dose 20 mg/kg	1	48	Risk Ratio (M‐H, Fixed, 95% CI)	1.1 [0.74, 1.63]
3 Failure to have clinical response at week 10	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
4 Adverse events	3	149	Risk Ratio (M‐H, Random, 95% CI)	0.85 [0.32, 2.31]
5 Serious adverse events	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
6 Withdrawal due to adverse events	3	149	Risk Ratio (M‐H, Fixed, 95% CI)	0.77 [0.36, 1.68]
6.1 Dose 10 mg/kg or 20 mg/kg	3	149	Risk Ratio (M‐H, Fixed, 95% CI)	0.77 [0.36, 1.68]

Outcome or subgroup title	No. of studies	Statistical method	Effect size
1 Failure to achieve clinical response at week 12	1	Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
2 Adverse events	1	Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
3 Withdrawal due to adverse events	1	Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to enter clinical remission at 6 to 10 weeks	7	773	Risk Ratio (M‐H, Fixed, 95% CI)	0.86 [0.76, 0.98]
2 Failure to achieve clinical response at week 10 or 12 or 14	5	617	Risk Ratio (M‐H, Fixed, 95% CI)	0.77 [0.64, 0.93]
3 Failure to maintain clinical remission at 52 weeks	2	155	Risk Ratio (M‐H, Random, 95% CI)	0.87 [0.52, 1.47]
4 Adverse events	9	852	Risk Ratio (M‐H, Fixed, 95% CI)	0.87 [0.75, 1.02]
5 Serious adverse events	3	520	Risk Ratio (M‐H, Fixed, 95% CI)	1.70 [0.29, 10.01]
6 Withdrawal due to adverse events	9	858	Risk Ratio (M‐H, Fixed, 95% CI)	0.86 [0.57, 1.29]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Failure to achieve clinical response at week 12	2	100	Risk Ratio (M‐H, Fixed, 95% CI)	0.57 [0.29, 1.10]
2 Adverse events	2	100	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.76, 1.12]
3 Withdrawal due to adverse events	2	100	Risk Ratio (M‐H, Fixed, 95% CI)	0.82 [0.19, 3.45]

Methods	Randomized, placebo‐controlled trial Trial duration was 6 weeks
Participants	Patients aged 15‐70 with documented histological or radiological diagnosis of CD on stable therapy for 28 days prior to study entry (N = 72)
Interventions	Oral cotrimoxazole 960 mg twice daily (BID) with oral placebo (n = 16) Oral metronidazole 400 mg BID with oral placebo (n = 18) Oral cotrimoxazole 960 mg BID and oral metronidazole 400 mg BID (n = 21) Oral placebo (n = 17)
Outcomes	Primary outcome: improvement in clinical score (defined by the authors: general health yesterday, abdominal pain yesterday, number of liquid stools yesterday, abdominal mass, complications) Investigators also looked at changes in fecal flora and hematologic measures
Notes	Patients were 15 years and older Trial did not mention whether it was blinded Intention to treat not performed
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias)   All outcomes	Unclear risk	Not described
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Not described
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	16 patients did not complete the study More patients in the cotrimoxazole and metronidazole group (8/21) failed to complete the trial compared to the other groups (placebo: 4/17; cotrimoxazole: 4/16; metronidazole 4/18) Study did not quantify adverse events based on treatment group (7/72)
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported
Other bias	Low risk	No other apparent sources of bias

Methods	Randomized active comparator trial. Study participants and investigators were not blinded Intention to treat not performed 6 week trial
Participants	Patients aged >18 with acute ileal or colonic CD with no need for steroid treatment and a CDAI between 150 and 300 (N = 40)
Interventions	Oral ciprofloxacin 500 mg BID (n = 18) or oral mesalazine 2g BID (n = 22) for 6 weeks
Outcomes	Primary outcome: Clinical remission (CDAI ≤ 150) at 6 weeks Secondary outcomes: Partial remission (CDAI ≤ 150 but change in CDAI between 50‐75), treatment failure (increase in CDAI value or insufficient improvement at week 3 with change in CDAI > 50 or absence of complete or partial remission at week 6)
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not discussed
Allocation concealment (selection bias)	Unclear risk	Not discussed
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Trial was not blinded
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Trial was not blinded
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Two patients stopped treatment in ciprofloxacin group (one for intolerance, one for personal reasons) One patient was lost to follow up in mesalamine group
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported
Other bias	Low risk	No other apparent sources of bias

PERMALINK

Antibiotics for induction and maintenance of remission in Crohn's disease

Cassandra M Townsend

Claire E Parker

John K MacDonald

Tran M Nguyen

Vipul Jairath

Brian G Feagan

Reena Khanna

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Results

Description of studies

Results of the search

1.

Included studies

Excluded studies

Risk of bias in included studies

2.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

Summary of findings for the main comparison. Antibiotic compared to placebo for induction and maintenance of remission in Crohn's disease.

Summary of findings 2. Antibiotic with anti‐TNF compared to placebo with anti‐TNF for induction of remission in Crohn's disease.

1. Ciprofloxacin (500 mg twice daily) compared to placebo for induction of remission in Crohn's disease.

2. Rifaximin (800 mg to 2400 mg daily) compared to placebo for induction of remission in Crohn's disease.

3. Metronidazole (400 mg to 500 mg twice daily) compared to placebo for induction of remission in Crohn's disease.

4. Clarithromycin (1 g/day) compared to placebo for induction of remission in Crohn's disease.

5. Cotrimoxazole (960 mg twice daily) compared to placebo for induction of remission in Crohn's disease.

6. Ciprofloxacin (500 mg twice daily) and metronidazole (250 mg four times daily) compared to methylprednisolone (0.7‐1 mg/kg daily) for induction and maintenance of remission in Crohn's disease.

7. Ciprofloxacin (500 mg twice daily) and metronidazole (500 mg twice daily) and budesonide (9 mg/daily) compared to placebo with budesonide (9 mg/daily) for induction of remission in Crohn's disease.

8. Ciprofloxacin (500 mg twice daily) compared to mesalazine (2 g twice daily) for induction of remission in Crohn's disease.

9. Ciprofloxacin (500 mg twice daily) with adalimumab compared to placebo with adalimumab for induction and maintenance of remission in Crohn's disease.

10. Ciprofloxacin (500 mg twice daily) with infliximab compared to placebo with infliximab for induction of remission in Crohn's disease.

11. Clarithromycin and antimycobacterial compared to placebo for maintenance of remission in Crohn's disease.

12. Metronidazole and cotrimoxazole compared to placebo for induction of remission in Crohn's disease.

Discussion

Summary of main results

Overall completeness and applicability of evidence

Quality of the evidence

Methods	Multicenter, randomised, double‐ blind, placebo controlled trial Modified intention to treat Study duration was 24 weeks
Participants	Patients aged 18‐70 with active CD and perianal activity (N = 76)
Interventions	Patient were randomised to oral ciprofloxacin 500 mg BID (n = 37) or placebo (n = 39) for 12 weeks and followed for a total of 24 weeks All patients were treated with self‐administered adalimumab (patients were given induction dosing of 160 mg at day 0 and 80 mg at week 2, followed by maintenance of 40 mg every 4 weeks until week 24)
Outcomes	Primary outcome: Clinical response defined as a 50% reduction in perianal fistulas Secondary outcomes: Clinical remission, improvement in CDAI, Pediatric Crohn's Disease Activity Index (PCDAI) and Inflammatory Bowel Disease Questionnaire (IBDQ)
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed through a centralised randomisation schedule in a 1:1 ratio
Allocation concealment (selection bias)	Low risk	Centralised randomisation was performed
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Double‐blind trial
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Clinical exam was performed at week 0, 12 and 24 by physician blinded for treatment allocation. Physician counted number of draining fistulas and excluded presence of abscesses
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Three patients from each treatment group were excluded from analysis. Reasons included (4 had no fistulas at baseline, 2 had a drain removed that was not according to protocol)
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported
Other bias	Low risk	No other apparent sources of bias Note: Ciprofloxacin group lost 3 patients after initial randomisation (34/37). Placebo group lost 3 patients after initial randomisation (36/39). Reasons reported

Methods	Randomized, placebo‐controlled trial. Investigators were blinded. Intention to treat Twelve week trial
Participants	Patients age 18 and older with active CD (CDAI > 200 and CRP ≥ 10 mg/L) (N = 41)
Interventions	Patients were randomised to oral clarithromycin 1g daily (n = 19) or oral placebo (n = 22) for 3 months
Outcomes	Primary outcome: Clinical remission (CDAI ≤ 150) or clinical response (fall in CDAI by ≥70 from pre‐treatment level) at 3 months Secondary outcomes: Decrease in van Hees Activity Index, remission as per Harvey Bradshaw Index (≤ 4), improvement in IBDQ and decrease in serum CRP
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization was carried out by pharmacist independent of the trial. Randomization was performed by a random allocation sequence in blocks of four (two active and two placebo)
Allocation concealment (selection bias)	Low risk	Centralized randomization by pharmacy
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Patients and investigators were blinded to treatment allocation Placebo was identical in size, colour and taste
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Investigators were blinded to randomisation assignments
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	More patients withdrawn from placebo group (7/19 from clarithromycin group, 12/22 of placebo group). More patients had worsening baseline disease in placebo group (4 in treatment group versus 9 in placebo group)
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported
Other bias	Low risk	No other apparent sources of bias

Methods	Randomized active comparator trial, some investigators were blinded. Investigators who were evaluating clinical status and results were blinded Investigators who were administering study medication or monitoring compliance were not blinded. Patients were not blinded Intention to treat not performed Twelve week trial If achieved remission continued therapy for 1 year
Participants	Patients with current active colonic or ileal CD with CDAI > 200 (N = 41)
Interventions	Oral ciprofloxacin 500 mg BID and metronidazole 250 mg QID (n = 22) or methylprednisolone 0.7‐1 mg/kg/day with a variable taper to 40 mg daily, followed by a taper of 4 mg weekly (n = 19)
Outcomes	Primary outcome: Clinical remission defined by CDAI ≤ 150 Secondary outcomes: Recovery of abdominal pain, diarrhea, fever, extraintestinal disease, improvement of general well being and weight gain
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization performed using a random number table
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias)   All outcomes	High risk	Physicians who supervised randomisation and monitored compliance and side effects were aware of treatment assignment All other investigators were blinded
Blinding of outcome assessment (detection bias)   All outcomes	High risk	Physicians who supervised randomisation and monitored compliance and side effects were unblinded
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Many more patients discontinued in the antibiotic group (11/22) compared to steroid group (5/19) Specific reasons for withdrawal not reported
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported
Other bias	Low risk	No other apparent sources of bias

Methods	Randomized placebo controlled trial. Investigators and patients were blinded Intention to treat Twelve week trial
Participants	Patients aged 18‐70 with mild to moderate CD (CDAI 200‐400) (N = 83)
Interventions	Randomized to receive oral rifaximin 800 mg daily (n = 25) and oral placebo, oral rifaximin 800 mg BID (n = 29) or oral placebo BID (n = 29). Other immunosuppressant’s kept at same dose of during study period
Outcomes	Primary outcome: Clinical remission defined by CDAI ≤ 150 Secondary outcomes: Reduction of CDAI ≥ 70 points from baseline, treatment failure and changes in quality of life (Inflammatory Bowel Disease questionnaire)
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Blinding of participants and investigators
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Investigators blinded to participant assignment. CDAI and the Inflammatory Bowel Disease Questionnaire (IBDQ) scores were calculated by blinded investigators
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Discontinuation and reasons for withdrawal were fairly even between each group. (rifaximin 800 mg daily and placebo 8/25, rifaximin 800 mg BID 12/29 or placebo BID 13/29)
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported
Other bias	Low risk	No other apparent sources of bias

Methods	Randomized, double blind, placebo controlled Intention to treat not performed Twenty‐four week trial
Participants	Patients aged 18‐75, active ileal or colonic CD (CDAI 220‐400) (N = 402)
Interventions	Randomized to oral rifaximin 400 mg BID (n = 104), oral rifaximin 800 mg BID (n = 98), oral rifaximin 1200 mg BID (n = 99) or oral placebo (n = 101)
Outcomes	Primary outcome: Clinical remission (CDAI < 150) at 12 weeks Secondary outcomes: Number of patients who achieved clinical response at week 12 (reduction in CDAI score of 100 points), number of patients who maintained clinical remission at week 14 and 24 and number of patients with treatment failure (failure to achieve loss of at least 70 in CDAI score after 1 month, or an increase in CDAI score of >100 points from the baseline)
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computerized randomisation procedure
Allocation concealment (selection bias)	Low risk	Centralized randomization using an Interactive Voice Response System
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Double‐blind
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Questionaires used to collect outcomes
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	High rate of withdrawal from the study in each treatment group. (rifaximin 400 mg BID (41/106), rifaximin 800 mg BID (33/99), rifaximin 1200 mg BID (41/103) or placebo (43/102)
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported
Other bias	Low risk	No other apparent sources of bias

Methods	Randomized, placebo controlled trial with blinded investigators and patients Intention to treat not performed Induction phase 16 weeks in duration, maintenance phase 104 weeks in duration, trial 156 weeks total; all patients were started on a 16‐week tapering course of prednisolone starting at 40 mg/day
Participants	Patients over age 18 with active CD (CDAI >200) (N = 213 for induction phase and N = 122 for maintenance phase)
Interventions	Randomized to oral clarithromycin 750 mg/day, oral rifabutin 450 mg/ day, oral clofazimine 50 mg/day (n = 102) or oral placebo (three placebo pills were given to placebo group) (n = 111) Maintenance phase: antibiotics (n = 67), placebo (n = 55)
Outcomes	Primary outcomes: The number of patients who had at least 1 relapse by 52, 104 or 156 weeks Secondary outcomes: Clinical remission at week 16 (CDAI ≤ 150), relapses within each study phase, time to first relapse, adverse events, endoscopic remission, need for surgery and quality of life
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Double blind: matched placebo
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Investigators and participants were blinded to treatment allocation. Investigators were more likely to identify those on active treatment
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Ninety‐one subjects were withdrawn during the induction phase (56 placebo group, 35 antibiotic group). Reasons included
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported
Other bias	Low risk	No other apparent sources of bias

Methods	Randomized, placebo controlled with blinded investigators Intention to treat not performed Eight week trial
Participants	Patients aged 14 years or older with active CD (CDAI 200‐400) (N = 134)
Interventions	Patients were randomised to receive oral ciprofloxacin 500 mg BID and oral metronidazole 500 mg BID (n = 66) or oral placebo (n = 68) All patients received oral budesonide 9 mg once daily
Outcomes	Primary outcome:The number of patients in remission at week 8, defined by CDAI < 150 Secondary outcomes: The mean changes in scores on the CDAI, quality of life, and adverse drug reactions
Notes	Patients were 14 years and older
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer generated randomisation
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Investigators and participants were blinded to treatment assignment Double blind: identical placebo
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Investigators were unaware of treatment assignments
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Sixteen patients in the placebo group and 22 patients in the antibiotic group did not complete the treatment course
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported
Other bias	Low risk	No other apparent sources of bias

Methods	Randomized, double‐blind, placebo controlled Intention to treat Sixteen week trial
Participants	N = 105, Patients with active CD with CDAI between 180‐450 (age not mentioned)
Interventions	Patients were randomised to oral metronidazole 20 mg/kg (n = 33) or oral metronidazole 10 mg/kg (n = 30) or oral placebo (n = 36)
Outcomes	Primary outcome: Clinical remission (CDAI < 150) at week 16 Secondary outcomes: Changes in plasma orosomucoid and C reactive protein levels Authors also looked at the effect of metronidazole on disease location
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Both patients and investigators were blinded to the treatment assignment
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Both patients and investigators were blinded to the treatment assignment. No clinical measures were part of outcomes
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Equal discontinuation rate among study groups, metronidazole 10 mg/kg 11/33, metronidazole 20 mg/kg 9/30. But higher attrition rate in placebo group 20/36
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported
Other bias	Low risk	No other apparent sources of bias

Methods	Randomized placebo controlled trial. Blinded participants and investigators Intention to treat Ten week trial
Participants	Patients aged 16 and older with confirmed CD with 1 or more actively draining perianal fistulas (N = 25)
Interventions	Patients were randomised to receive oral ciprofloxacin 500 mg BID (n = 10), oral metronidazole 500 mg BID (n = 7) or oral placebo (n = 8)
Outcomes	Primary outcome: Closure of fistulas, meaning that they were no longer draining Secondary outcomes:Clinical remission at weeks 2 and 6, improvement in fistula symptoms at weeks 2, 6, and 10 (defined as 50% reduction in draining fistulas from baseline), short‐term durability of fistula closure defined as maintaining remission for at least 4 weeks, patient and physician global assess‐ ment measured at weeks 2, 6, and 10, PDAI at weeks 2, 6, and 10, IBDQ at weeks 2, 6, and 10; and CDAI at weeks 2, 6, and 10
Notes	Patients 16 years and older
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization was done using computer randomisation
Allocation concealment (selection bias)	Low risk	Medication distrubution was done by the investigational pharmacist service
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Both patients and investigators were blinded to treatment assignments Patients took tablets with similar appearance
Blinding of outcome assessment (detection bias)   All outcomes	Unclear risk	Specific efforts to maintain blinding among study physicians not described
Incomplete outcome data (attrition bias)   All outcomes	Unclear risk	Seven subjects had left the study early, more in the metronidazole group (5/7) compared to others (1/10 in ciprofloxacin group and 1/8 in placebo group) Reasons for discontinuation included lost to follow‐up, withdrawal of consent, adverse events, and other reasons
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported
Other bias	Low risk	No other apparent sources of bias

Methods	Randomized controlled trial with blinded investigators and participants Intention to treat Eighteen week trial
Participants	Patients aged with a diagnosis of CD that was complicated by at least one perianal fistula (N = 24)
Interventions	Patients were randomised to receive oral ciprofloxacin 500 mg twice daily (n = 11) or oral placebo (n = 13) All patients also received 5 mg/kg infliximab at weeks 6, 8, and 12. Three patients received intravenous hydrocortisone prior to infliximab infusion
Outcomes	Primary outcome: Clinical response (50% or greater reduction in draining fistulae confirmed by no drainage despite firm finger compression) Secondary outcomes: PDAI and 3D endoscopic ultrasound with 3% hydrogen peroxide used as a contrast medium
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Method of randomisation not described
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding of participants and personnel (performance bias)   All outcomes	Low risk	Both patients and investigators were blinded to treatment assignments
Blinding of outcome assessment (detection bias)   All outcomes	Low risk	Both patients and investigators were blinded to treatment assignments
Incomplete outcome data (attrition bias)   All outcomes	Low risk	Rate of discontinuation was similar in both treatment groups 2/11 in ciprofloxacin group and 2/13 in placebo group
Selective reporting (reporting bias)	Low risk	All expected outcomes were reported
Other bias	Low risk	No other apparent sources of bias