Comparative efficacy of treatments for Clostridium difficile infection: a systematic review and network meta-analysis

Articles

Comparative efficacy of treatments for Clostridium difficile infection: a systematic review and network meta-analysis Tumas Beinortas*, Nicholas E Burr*, Mark H Wilcox, Venkataraman Subramanian

Summary

Background Several new treatments for Clostridium difficile infections have been investigated. We aimed to compare and rank treatments for non-multiply recurrent infections with C difficile in adults. Methods We did a random effects network meta-analysis within a frequentist setting to obtain direct and indirect comparisons of trials. We searched MEDLINE, Embase, Web of Science, Cochrane Central Register of Controlled Trials, and ClinicalTrials.gov for published and unpublished trials from the creation of these databases until June 30, 2017. We included randomised controlled trials of treatments for non-multiply recurrent infections with confirmed C difficile in adults (at least 18 years) that reported both primary cure and recurrence rates, and we used the Cochrane Risk of Bias tool to appraise trial methods. For our analysis, we extracted the total numbers of patients with primary cure and recurrence from published and unpublished reports. The primary outcome was sustained symptomatic cure, defined as the number of patients with resolution of diarrhoea minus the number with recurrence or death. Findings Of 23 004 studies screened, 24 trials, which comprised 5361 patients and 13 different treatments, were included in the analysis. The overall quality of evidence was rated as moderate to low. For sustained symptomatic cure, fidaxomicin (odds ratio 0·67, 95% CI 0·55–0·82) and teicoplanin (0·37, 0·14–0·94) were significantly better than vancomycin. Teicoplanin (0·27, 0·10–0·70), ridinilazole (0·41, 0·19–0·88), fidaxomicin (0·49, 0·35–0·68), surotomycin (0·66, 0·45–0·97), and vancomycin (0·73, 0·56–0·95) were better than metronidazole. Bacitracin was inferior to teicoplanin (0·22, 0·06–0·77) and fidaxomicin (0·40, 0·17–0·94), and tolevamer was inferior to all drugs except for LFF571 (0·50, 0·18–1·39) and bacitracin (0·67, 0·28–1·58). Global heterogeneity of the entire network was low (Cochran’s Q=15·70; p=0·47). Interpretation Among the treatments for non-multiply recurrent infections by C difficile, the highest quality evidence indicates that fidaxomicin provides a sustained symptomatic cure most frequently. Fidaxomicin is a better treatment option than vancomycin for all patients except those with severe infections with C difficile and could be considered as a first-line therapy. Metronidazole should not be recommended for treatment of C difficile.

Lancet Infect Dis 2017 Published Online July 16, 2018 http://dx.doi.org/10.1016/ S1473-3099(18)30285-8 See Online/Comment http://dx.doi.org/10.1016/ S1473-3099(18)30308-6 *Contributed equally Department of Gastroenterology (T Beinortas MBBCh, N E Burr MBBS, V Subramanian FRCP) and Department of Microbiology (Prof M H Wilcox MD), Leeds Teaching Hospitals NHS Trust; and Leeds Institute of Biomedical and Clinical Sciences, University of Leeds, Leeds, UK (N E Burr, V Subramanian) Correspondence to: Dr Venkataraman Subramanian, Department of Gastroenterology, St James University Hospital, Leeds LS9 7TF, UK [email protected]

Funding None. Copyright © 2018 Elsevier Ltd. All rights reserved.

Introduction Reclassification of Clostridium difficile as Clostridioides difficile has been proposed,1 but a preference for the established name prevails. C difficile infections are increasing in incidence and are the most common health care-associated infections in the USA; they are also increasing in incidence in low-income and middle-income countries.2,3 In the USA, there were 29 000 deaths due to infections with C difficile in 2011 and, in 2014, they led to a financial burden of US$5·4 billion.4–6 For more than three decades, metronidazole and vancomycin have been the principal treatment options for C difficile infections. However, sub-optimal numbers of sustained cures and the increasing prevalence and associated morbidity and mortality from infections with C difficile have warranted the development and evaluation of new therapeutic drugs. After showing a higher number of sustained clinical cures than vancomycin,7 fidaxomicin was approved for treatment of infections with C difficile in 2011. However, a long-term response was not achieved in 29% of patients,

and research to develop several drugs to provide a lasting cure are ongoing.8,9 Clinical trials have evaluated many treatments for infections with C difficile, such as tolevamer, an orally administered toxin-binding polymer, and several antimicrobial drugs, such as bacitracin, fusidic acid, surotomycin, ridinidazole, teicoplanin, LFF571, nitazoxanide, cadazolid, and rifaximin. Several pairwise comparison meta-analyses10–14 have investigated the efficacy of treatments for infections with C difficile. However, these studies have mostly focused on a subset of treatments that were investigated for C difficile infection. Additionally, there have been several novel and unpublished trials which, to our knowledge, have not yet been included or synthesised in a systematic review. Furthermore, most of the agents do not have direct trial comparisons, making it impossible to generate a hierarchy of treatments through pairwise meta-analyses. We therefore did a network meta-analysis, in which we aimed to compare and rank treatments for non-recurrent infections with C difficile in adults.

www.thelancet.com/infection Published online July 16, 2018 http://dx.doi.org/10.1016/S1473-3099(18)30285-8

1

Articles

Research in context Evidence before this study We did a systematic literature search of PubMed, Embase, and Web of Science for systematic reviews and meta-analyses of treatments for Clostridium difficile infections. We searched for articles published in any language between Jan 1, 2010, and June 1, 2017, with the search terms “Clostridium difficile”, “meta-analysis”, “CDI”, “CDAD”, “systematic review”, and “meta analysis” in all fields, but we restricted the search to meta-analyses and systematic reviews. Only meta-analyses of randomised controlled trials for treatment of C difficile infections were included. 418 records were identified, of which four articles met the inclusion criteria. One meta-analysis made a direct comparison of fidaxomicin, metronidazole, and vancomycin, one meta-analysis examined fidaxomicin and vancomycin only, and two meta-analyses evaluated all antibiotics trialled for treatment of C difficile. We found no network meta-analyses. The most comprehensive Cochrane meta-analysis by Nelson and colleagues, published in 2017, used pairwise comparisons for different antibiotics when direct evidence was available. There have been no analyses of indirect evidence of treatments for initial infections with C difficile that would enable ranking the treatments in order of efficacy.

Methods

Search strategy and selection criteria

See Online for appendix

2

We searched MEDLINE, Embase, Web of Science, Cochrane Central Register of Controlled Trials, and ClinicalTrials.gov up until June 30, 2017, for full texts, conference abstracts, and proceedings that described therapeutic randomised controlled trials (RCTs) against infections with C difficile. We also searched the reference lists of systematic reviews of treatments for C difficile that were published between Jan 1, 2012, and June 30, 2017. To maximise the number of papers found, both MeSH and free text terms were used. The search terms are listed in the appendix. No language restrictions were applied, and non-English articles were translated. We also searched the databases of pharmaceutical companies and contacted study authors or the pharmaceutical companies (when authors were not listed) of registered but unpublished trials. Two authors (TB, NEB) independently reviewed and assessed the eligibility of titles, abstracts, and studies that were deemed relevant for review of their full texts. Any discrepancies were resolved through discussion with another author (VS). A systematic review and network meta-analysis was done in accordance with the Preferred Reporting Items for Systematic Reviews and Network Meta-Analyses (PRISMA) checklist.15 The appendix details the study protocol. TB and NEB reviewed all RCTs that investigated the therapeutic effects of at least two different treatments for infections with C difficile in full. Studies that investigated pharmacological treatments, probiotics, immunotherapy,

Added value of this study To our knowledge, this is the first network meta-analysis of pharmacological treatments for infections with C difficile. This meta-analysis evaluates 13 different treatments and allows comparison and ranking of efficacy for treatments that were not directly compared. We included three trials that have not been published and were not included in previous pairwise meta-analyses. Our study suggests that fidaxomicin is the treatment with the strongest evidence for providing a sustained symptomatic cure in patients with C difficile and that metronidazole is worse than several other drugs at giving a sustained symptomatic cure. We also show that teicoplanin and ridinidazole could potentially be effective treatments for these infections; however, their routine implementation should await results from larger trials. Implications of all the available evidence Our findings indicate that fidaxomicin and vancomycin can be recommended as first-line treatments for C difficile. Metronidazole cannot be recommended for treatment of C difficile. In Europe, if fidaxomicin or vancomycin are unavailable, treatment with oral teicoplanin might be a viable alternative.

and faecal microbiota transfer treatments were included if they met the inclusion criteria. Patients had to be adults (aged at least 18 years) with confirmed infections with C difficile (defined as active diarrhoea and a positive C difficile nucleic acid amplification test, or a positive C difficile cytotoxin assay result, or a stool culture that showed C difficile, or pseudomembranes when assessed with colonoscopy). Studies of patients who had multiple recurrence or relapse of C difficile (which comprises a minority of patients with infections with C difficile, and has a different prognosis from the overall patient cohort with a C difficile infection) were excluded from the analysis. The studies also had to report both the primary symptomatic cure and recurrence of diarrhoea. Studies were excluded if the data for the intention-to-treat analysis were not available, if they investigated the prophylactic effects of the drug (as opposed to the therapeutic effects), or if several drugs were simul­ taneously used to treat the C difficile infection.

Outcomes Our primary outcome was sustained symptomatic cure, which was calculated as the number of patients with a primary cure (resolution of diarrhoea, as defined by individual trial criteria) at the end of treatment, minus the number of patients with recurrence (recurrence of diarrhoea or requirement for additional treatment) or who died during the follow-up period. Secondary outcomes were primary cure (resolution of diarrhoea, per individual trial criteria) and recurrence rate (recurrence of diarrhoea or death within the follow-up period of each trial).

www.thelancet.com/infection Published online July 16, 2018 http://dx.doi.org/10.1016/S1473-3099(18)30285-8

Articles

Data analysis Two authors (TB and NEB) independently reviewed papers included in the final analyses and extracted relevant data (appendix) directly from the papers. The Cochrane risk of bias tool16 was used to assess the risk of bias (appendix) and Revman version 5.0 to generate the risk of bias tables. We analysed by intention to treat, in which dropouts were assumed to be treatment failures. Any discrepancies with data extraction or assessment for risk of bias were resolved through consensus decision with another author (VS). Multiple pairwise meta-analyses of antibiotics that are used to treat C difficile have been reported by Nelson and colleagues10 and were not repeated here. Network metaanalyses allow the comparison of evidence from clinical studies where directly comparative data are not available, and enables the ranking of treatments in order of efficacy.17 We did a random-effects network meta-analysis by use of a frequentist setting.18 We used the Netmeta package for R for numerical data analysis. A randomeffects model was used to obtain the relative treatment effects. Given its widespread use, vancomycin was chosen as a reference treatment. Forest plots were generated to illustrate the treatment effects relative to vancomycin. League tables were used to display the relative efficacy of all available pairwise comparisons of available treatments. The P score was used to rank treatments, which can have a value between 0 and 1; higher P scores indicate a greater chance of being the best treatment.19 A scatter plot was used to spatially visualise the partial order of treatments with regard to primary cure and recurrence rates. NetmetaXL version 1.6.1 was used to generate network graphs, which were used to illustrate the evidence base.20 Treatment estimates are presented as odds ratios (ORs) with 95% CIs. We did three prespecified sensitivity analyses. First, non-blinded studies were excluded because resolution of diarrhoea is a semi-objective outcome that can be adversely affected by absence of blinding. In a second sensitivity analysis, we excluded trials published before 2000 because the incidence of infections with C difficile has markedly increased since 2000, coinciding with the emergence of the hypervirulent BI/NAP1/027 strain.21 Finally, we also excluded studies with fewer than 50 participants in each study group, to test for smallstudy effects. In one post-hoc sensitivity analysis, we also excluded RCTs done before 1990. We did additional prespecified subgroup analyses and individual network meta-analyses for patients with severe and mild to moderate infections with C difficile, first and singly recurrent infections with C difficile, and patients aged younger than 65 years and 65 years or older. We stratified patients into different severity categories, as defined by each trial. These assessment criteria are summarised in the appendix. For subgroup analyses of fidaxomicin trials, we used review data14 because primary

publications did not provide the recurrence rate. Insufficient data were available to perform inpatient versus outpatient subgroup analyses. A generalised Cochran’s Q statistic was used to assess the homogeneity of the multivariate meta-analysis.22 To identify single-design and between-design contributions to global heterogeneity in the random effects model, the global Cochran’s Q score was further split into withindesigns22 and between-designs heterogeneity scores.23 The between-designs Q score was calculated on the basis of a full design-by-treatment interaction random effects model,24 defined with a generalised methods-of-moments estimate of the between-studies variance (ie, τ²).25 A network heat plot was used to visualise and identify the nodes of single-design inconsistency.22 We checked the consistency between direct and indirect evidence by use of so-called node-splitting.23 A p value of less than 0·10 was considered to be significant in inconsistency assessments. Comparison-adjusted funnel plots were generated with STATA (version 14.0) to assess publication and small-study bias.

Role of the funding source

For RevMan 5 see http://community.cochrane.org/ help/tools-and-software/ revman-5

For information on the Netameta package see https://www.rdocumentation. org/packages/netmeta/ versions/0.9-8

The funder of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Results We identified 29 976 references, of which 6972 were duplicates (figure 1). Of the 144 full-text articles reviewed, 19 publications (which described 20 RCTs) were deemed eligible and were included in the final network metaanalysis. Two additional unpublished RCTs26,27 were retrieved from pharmaceutical company databases, one of which was published 9 months after the search.26 Two additional unpublished RCTs were provided by direct communication with pharmaceutical companies and authors. One trial was in Japanese and all others were in English. Therefore, 24 RCTs that involved 5361 independent patients were included in the network meta-analysis. The studies that were included were published between 1983 and 2017, and investigated 13 pharmacological interventions against infections with C difficile (table 1).8,9,26–46 Follow up-time was between 21 and 30 days for all studies except those by Louie and colleagues,41 who reported outcomes at 56 days, and Guery and colleagues,26 who reported outcomes at 90 days. Guery and co​lleagues26 also reported results at 30 days of follow-up, which were used in our analysis to allow better comparison with the other studies. None of the faecal microbiota transfer, probiotic, or immunotherapy trials met the inclusion criteria. All included trials had an active control. The network was well balanced and interconnected: five treatments had more than 400 patients and there

www.thelancet.com/infection Published online July 16, 2018 http://dx.doi.org/10.1016/S1473-3099(18)30285-8

3

Articles

29 976 references identified by database search

6972 duplicates excluded

23 004 titles screened

21 445 titles excluded

1559 abstracts reviewed

1415 abstracts excluded

144 full-text articles reviewed

2 unpublished randomised controlled trials from pharmaceutical company databases

125 full-text articles excluded 44 duplicate reports of the same data 24 only included patients with recurrent Clostridium difficile infections 22 cohort studies 10 single-arm studies 5 preventive effect studies 5 case reports or case series 4 only compared different doses of the same drug 3 only used dual-active treatment 3 reviews 1 used healthy subjects 1 included patients with several causes of diarrhoea after operations 1 used a placebo group from another trial 1 missing data for an intention-to-treat-analysis 1 study report not available

2 unpublished RCTs from data inquiries to authors and pharmaceutical companies

19 articles (describing 20 RCTs) and 4 unpublished RCTs included in the network meta-analysis

Figure 1: Study selection

were 11 loops (in which every treatment in the loop has direct trial comparisons with at least two other drugs). The mean study sample size was 223 participants (range 12–629; table 1). Use of vancomycin was the most frequent intervention and was investigated in 21 RCTs (n=2107 participants). The next most common inter­ ventions were use of metronidazole (seven RCTs; n=563 participants) and fidaxomicin (six RCTs; n=881). The mean participant age was 63 years and 53% were female (table 1). The duration of treatment ranged from 4 to 25 days and the median duration of follow-up was 28 days (range 21–90). 17 (71%) trials were sponsored by industry, two (8%) jointly by government and industry and, for four (17%) trials, funding information was not 4

provided. Most RCTs were done in the USA, Canada, Australia, or Europe. The RCT by Mikamo and colleagues27 from 2016 was done in Japan, and Boix and colleagues38 also recruited patients from two centres in the Middle East. Ten (42%) trials were multinational. The overall quality of studies was moderate to low (figure 2; appendix). Random sequence generation procedures were adequate and clearly described in only ten (42%) RCTs, and seven (29%) of 24 RCTs were nonblinded. The network for efficacy assessment of sustained symptomatic cure can be seen in figure 3. Network graphs for primary cure and recurrence were identical. All drugs had at least one direct comparison with vancomycin. A summary of the pairwise comparisons is shown in figure 4. Teicoplanin (OR 0·37, 95% CI 0·14–0·94) and fidaxomicin (0·67, 0·55–0·82) were significantly better than vancomycin in attaining a sustained symptomatic cure. Vancomycin was superior to metronidazole (0·73, 0·56–0·95). Teicoplanin, ridini­ dazole, fidaxomicin, and surotomycin were also more efficacious than metronidazole (figure 4). Tolevamer was significantly inferior to all drugs except LFF571 and bacitracin. In our GRADE assessment, only fidaxomicin had a high-confidence treatment effect (appendix). Confidence in the treatment effects of teicoplanin was very low and confidence in the effects of ridinidazole was moderate. Vancomycin ranked seventh and metro­ nidazole ranked 11th among the 13 assessed drugs. No treatment was significantly superior to vancomycin in giving a primary symptomatic cure (appendix). Tolevamer was inferior to all treatments and metronidazole was inferior to vancomycin. Fidaxomicin had signifi-​ cantly fewer associated recurrences than vancomycin and metronidazole (appendix). Regarding number of re-​ currences, vancomycin ranked ninth and metronidazole ranked 11th among the 13 treatments. Heterogeneity for the entire network meta-analysis for sustained symptomatic cure, was not significant (Cochran’s Q, 15·70; p=0·47; τ², 0). Between-designs hetero­­geneity for sustained symptomatic cure was low (3·19; p=0·87) and non-significant for all 11 loops (appendix). Within-designs hetero­geneity (12·61; p=0·18) was higher because of significant pairwise vancomycin– metronidazole com­parison hetero­geneity (3·94; p=0·047). This heterogeneity originated from a markedly higher sustained symptomatic cure rate in the metronidazole group shown in the non-blinded trial by Teasley and colleagues39 than in other trials that investigated metronidazole and vancomycin. In this trial, 1:1 randomisation resulted in a markedly lower number of participants in the metronidazole group than in the vancomycin group (45 vs 56). A heatplot identified only a few faint nodes of in-​ consistency between direct and indirect evidence (appendix). These highlighted metronidazole–fusidic acid and fusidic acid–teicoplanin interactions were

www.thelancet.com/infection Published online July 16, 2018 http://dx.doi.org/10.1016/S1473-3099(18)30285-8

Articles

Dosing information

Follow-up, Proportion days of female participants

Mean Proportion of age, years participants with severe Clostridium difficile infections

Location of study

Sponsorship

Louie et al8 (2011)

125 mg vancomycin oral capsules four times daily for 10 days, n=327; 200 mg fidaxomicin oral capsules twice daily for 10 days, n=302

28

56%

62

39%

USA, Canada

Industry

Cornely et al9 (2012)

125 mg vancomycin oral capsules four times daily for 10 days, n=265; 200 mg fidaxomicin oral capsules twice daily for 10 days, n=270

28

61%

63

24%

USA, Canada, Europe

Industry

Guery26 (2017)

125 mg vancomycin oral capsules four times daily for 10 days, n=181; 200 mg fidaxomicin oral capsules twice daily for 5 days, then once daily every 2 days for 20 days, n=183

90*

58%

75

27%

Europe, Turkey

Industry

Mikamo et al27 (2016)

125 mg vancomycin oral liquid four times daily for 10 days, n=109; 200 mg fidaxomicin oral capsules twice daily for 10 days, n=106

28

52%

75

22%

Japan

Industry

Zar et al28 (2007)

125 mg vancomycin oral liquid four times daily for 10 days, n=82; 250 mg metronidazole oral capsules four times daily for 10 days, n=90

21

41%

58

48%

USA

Not specified

Wenisch et al29 (1996) 500 mg vancomycin oral capsules three times daily for 10 days, n=31; 500 mg metronidazole oral capsules three times daily for 10 days, n=31; 400 mg teicoplanin oral liquid twice daily for 10 days, n=28; 500 mg fusidic acid oral capsules three times daily for 10 days, n=29

30

48%

42

Not specified

Austria

Not specified

Wullt et al30 (2004)

400 mg metronidazole oral capsules three times daily for 7 days, n=64; 250 mg fusidic acid oral capsules three times daily for 7 days, n=67

30

65%

58

Not specified

Sweden

Government and industry

Young et al31 (1985)

125 mg vancomycin oral capsules four times daily for 7 days, n=21; 20 000 units bacitracin oral capsules four times daily for 4 days n=21

28

NA

62

Not specified

Australia

Not specified

Vickers et al32 (2017)

125 mg vancomycin oral capsules four times daily for 10 days, n=50; 200 mg ridinidazole oral capsules twice daily for 10 days, n=50

30

66%

57

16%

USA

Industry

Mullane et al33 (2015)

125 mg vancomycin oral capsules four times daily for 10 days, n=26; 200 mg LFF571 oral capsules four times daily for 10 days, n=46

30

65%

58

20%

USA, Canada

Industry

Louie et al34 (2015)

125 mg vancomycin oral capsules four times daily for 10 days, n=22; 250, 500, or 1000 mg cadazolid oral liquid twice daily for 10 days, n=62

30

39%

51

9%

Canada, Germany, UK, USA

Industry

Musher et al35 (2006)

250 mg metronidazole oral capsules four times daily for 10 days, n=44; 500 mg nitazoxanide oral capsules twice daily for 7d or 10 days, n=98

21

24%

68

Not specified

USA

Industry

Musher et al36 (2009)

125 mg vancomycin oral capsules four times daily for 10 days, n=27; 500 mg nitazoxanide oral capsules twice daily for 10 days, n=23

21

34%

63

41%

USA

Industry

Dudley et al37 (1986)

500 mg vancomycin oral liquid four times daily for 10 days, n=31; 25 000 units bacitracin oral liquid four times daily for 10 days, n=31

NA

60%

69

Not specified

USA

Industry

Boix et al38 (2017)

125 mg vancomycin oral capsules four times daily for 10 days, n=298; 250 mg surotomycin oral capsules twice daily for 10 days, n=308

30

40%

61

34%

USA, Canada, Europe, Middle East

Industry

Teasley et al39 (1983)

500 mg vancomycin oral capsules four times daily for 10 days, n=56; 250 mg metronidazole oral capsules four times daily for 10 days, n=45

21

NA

65

Not specified

USA

Government and industry

Lee et al40 (2016)

125 mg vancomycin oral capsules four times daily for 10 days, n=70; 125 or 250 mg surotomycin oral capsules twice daily for 10 days, n=139

28

63%

NA

6%

USA, Canada

Industry

Louie et al41 (2006)

125 mg vancomycin oral capsules four times daily for 10 days, n=96; 3 g or 6 g tolevamer oral capsules three times daily for 14 days, n=190

56

55%

67

1%

USA, Canada, UK

Industry

Johnson et al42 (2014; 301)

125 mg vancomycin oral capsules four times daily for 10 days, n=140; 375 mg metronidazole oral capsules four times daily for 10 days, n=149; 3g tolevamer oral liquid three times daily for 14d, n=285

28

53%

62

34%

USA, Canada, Industry Europe, Canada

Johnson et al42 (2014; 302)

125 mg vancomycin oral capsules four times daily for 10 days, n=126; 375 mg metronidazole oral capsules four times daily for 10 days, n=140; 3 g tolevamer oral liquid three times daily for 14 days, n=278

28

54%

68

24%

USA, Canada, Europe

Industry

de Lalla et al43 (1992)

500 mg vancomycin oral liquid four times daily for 10 days, n=24; 100 mg teicoplanin oral liquid twice daily for 10 days, n=27

30

69%

NA

Not specified

Italy

Not specified

Thabit et al44 (2016)

125 mg vancomycin oral capsules four times daily for 10 days, n=5; 200 mg fidaxomicin oral capsules twice daily for 10 days, n=7

28

50%

70

Not specified

USA

Industry

Pardi et al45 (2012)

125 mg vancomycin oral capsules four times daily for 10 days, n=119; 400 mg rifaximin oral capsules three times daily for 10 days, n=119

28

61%

60

Not specified

USA

Industry

Mitra et al46 (2017)

200 mg ridinidazole oral capsules twice daily for 10 days, n=14; 200 mg fidaxomicin oral capsules twice daily for 10 days, n=13

30

NA

NA

UK

Industry

7%

Dosing information are dose and drug, form of drug, frequency and length of treatment, and number of participants for each treatment group. NA=not available. *Authors presented follow-up results up to 90 days, but we use 30-day follow-up results for our analysis to enable comparison between network meta-analysis studies.

Table 1: Summary of the included trials

www.thelancet.com/infection Published online July 16, 2018 http://dx.doi.org/10.1016/S1473-3099(18)30285-8

5

Zar,26 2007

Young,29 1985

Wullt,28 2004

Wenisch,27 1996

Vickers,30 2017

Thabit,44 2016

Teasley,39 1983

Mikamo,27 2016

Musher,36 2009

Musher,35 2006

Mullane,33 2015

Louie,34 2015

Louie,31 2011

Louie,41 2006

Lee,40 2016 (302)

Johnson,42 2014 (302)

Johnson,42 2014 (301)

Dudley,37 1986

de Lalla,43 1992

Low risk of bias Unknown risk of bias High risk of bias

Cornely,32 2012

Boix,38 2017

Articles

Random sequence generation (selection bias) Allocation concealment (selection bias) Masking of participants and personnel (performance bias) Masking during outcome assessment (detection bias) Incomplete outcome data (attrition bias) Selective reporting (reporting bias)

Figure 2: Summary of risk of bias assessment The Johnson et al study (2014) reported two trials: 301 and 302. Both were of identical design (appendix). Teicoplanin Surotomycin

Ridinidazole Tolevamer

Rifaximin Nitazoxanide

Vancomycin

LFF571

Bacitracin

Fusidic acid

Metrinidazole Cadazolid Fidaxomicin

Figure 3: Network of eligible comparisons for efficacy of treatments of Clostridium difficile Line width is proportional to the number of trials comparing every pair of treatments. The size of the circle is proportional to the number of patients assigned to receive the treatment.

affected by results derived from a four-arm, non-blinded RCT by Wenisch and colleagues.29 Wenisch and colleagues29 found a high sustained cure for teicoplanin and significantly higher recurrence rate for patients treated with fusidic acid than a subsequent moderate quality to high quality RCT by Wullt and colleagues30 that compared fusidic acid and metronidazole. Comparisons of direct versus indirect treatment estimates did not reveal any significant differences (appendix). A comparison-adjusted funnel plot did not show any small trial or publication bias (appendix). For 6

primary cure, global heterogeneity was low (Cochran’s Q, 13·52; p=0·63; τ², 0; appendix). For recurrence, global heterogeneity was significant (24·02; p=0·09; τ², 0·089), mainly due to significant between-design heterogeneity, which was present in nine of 11 loops (appendix). In isolation, recurrence network meta-analysis results should be interpreted with caution. Exclusion of non-blinded trials eliminated all teicoplanin and LFF571 RCTs from the network metaanalysis (appendix). Since they had similar P-scores, ridinilazole and fidaxomicin remained the top-ranking treatments. Estimates of other effect sizes did not change significantly after exclusion of non-blinded trials and global heterogeneity was low (Cochran’s Q, 7·97; p=0·44; τ², 0). Ridinilazole and fidaxomicin were also the highest ranked drugs after exclusion of small studies (<50 patients in each group) and RCTs published before 2000. Due to low total participant numbers in ridinilazole treatment group (n=64 participants), confidence intervals of its treatment effect estimates were very wide. All sensitivity analyses resulted only in minimal changes in estimates of treatment effect from those seen in the overall network meta-analysis (appendix). Few trials had available data for subgroup evaluation and there were no subgroup data for bacitracin, teicoplanin, rifaximin, LFF571, or cadazolid. In subgroup analyses, fidaxomicin was superior to vancomycin in mild to moderate infections with C difficile, initial and non-initial C difficile infection, and in both patients younger than 65 years and patients 65 years or older (table 2). Ridinilazole was significantly better than vancomycin in attaining a sustained symptomatic cure in patients with mild to moderate infections with C difficile and who were younger than 65 years. Ridinilazole ranked

www.thelancet.com/infection Published online July 16, 2018 http://dx.doi.org/10.1016/S1473-3099(18)30285-8

Articles

0·9386 TEIC

0·8280

0·65 (0·20–2·12)

Greatest chance of being the best treatment (higher P score)

RID

0·55 0·84 (0·21–1·44) (0·41–1·74)

P score Drug

0·7922 FID

0·6951

0·53 0·82 0·97 (0·13–2·15) (0·23–2·86) (0·34–2·78)

CAD

0·41 0·63 0·75 0·77 (0·15–1·10) (0·29–1·35) (0·53–1·06) (0·26–2·24)

0·5820 SUR

0·5405

0·39 0·60 0·72 0·74 0·96 (0·13–1·21) (0·23–1·58) (0·37–1·41) (0·22–2·49) (0·48–1·94)

NIT

0·4850

0·37 0·57 0·67 0·69 0·90 0·93 (0·14–0·94) (0·28–1·15) (0·55–0·82) (0·25–1·94) (0·68–1·19) (0·49–1·78) 0·34 (0·11–1·01)

VAN

0·4296

0·52 0·62 0·64 0·83 0·86 0·92 (0·21–1·28) (0·34–1·12) (0·20–2·06) (0·44–1·55) (0·37–2·02) (0·53–1·61)

0·31 0·48 0·57 0·59 0·77 0·80 0·85 (0·11–0·89) (0·19–1·23) (0·30–1·09) (0·18–1·95) (0·39–1·50) (0·35–1·84) (0·47–1·57)

RFX

0·3794

0·93 (0·41–2·11)

FUA

0·29 0·45 0·54 0·55 0·72 0·75 0·80 0·87 0·94 (0·08–1·15) (0·13–1·52) (0·20–1·46) (0·13–2·29) (0·26–1·99) (0·23–2·42) (0·30–2·13) (0·28–2·68) (0·30–2·97)

0·3635 LFF571

0·27 0·41 0·49 0·51 0·66 0·68 0·73 0·79 0·86 0·92 (0·10–0·70) (0·19–0·88) (0·35–0·68) (0·17–1·46) (0·45–0·97) (0·37–1·27) (0·56–0·95) (0·43–1·47) (0·48–1·52) (0·33–2·53)

Lower chance of being the best treatment (lower P score)

0·2411 MET

0·2006

0·22 0·34 0·40 0·42 0·54 0·56 0·60 0·65 0·70 0·75 0·82 (0·06–0·77) (0·11–1·00) (0·17–0·94) (0·11–1·55) (0·23–1·28) (0·20–1·59) (0·26–1·36) (0·24–1·76) (0·25–1·95) (0·21–2·70) (0·35–1·94)

BAC

0·0245

0·15 0·23 0·27 0·28 0·36 0·38 0·40 0·44 0·47 0·50 0·55 0·67 (0·06–0·39) (0·11–0·48) (0·20–0·37) (0·10–0·80) (0·25–0·53) (0·20–0·73) (0·32–0·51) (0·24–0·80) (0·25–0·87) (0·18–1·39) (0·42–0·72) (0·28–1·58)

TOL

Figure 4: League table of pairwise comparisons in network meta-analysis for attaining a sustained symptomatic cure Treatments are ordered in the rank of their chance of being the best treatment. Numbers in grey boxes are P scores, which are used to rank the treatments. Treatment estimates are provided as odds ratios with 95% CIs. Significant pairwise comparisons are highlighted. TEIC=teicoplanin. RID=ridinidazole. FID=fidaxomicin. CAD=cadazolid. SUR=surotomycin. NIT=nitazoxanide. VAN=vancomycin. RFX=rifaximin. FUA=fusidic acid. MET=metronidazole. BAC=bacitracin. TOL=tolevamer. Ridinidazole

Fidaxomicin

Nitazoxanide

Metronidazole

Surotomycin

Tolevamer

Fusidic acid

Severe Clostridium difficile infection

0·37 (0·05–3·06)

0·57 (0·30–1·11)

0·64 (0·09–4·37)

1·47 (0·78–2·78)

4·33 (0·14–137·06)

2·67 (1·30–5·49)*

NA

Mild to moderate C difficile infection

0·36 (0·14–0·93)*

0·47 (0·33–0·66)*

0·80 (0·15–4·26)

1·57 (1·06–2·32)*

0·59 (0·31–1·12)

2·86 (2·00–4·08)*

NA

Initial C difficile infection

0·43 (0·18–1·05)

0·52 (0·38–0·70)*

0·71 (0·18–2·76)

1·34 (0·90–1·99)

0·56 (0·28–1·11)

3·10 (2·18–4·40)*

0·84 (0·37–1·90)

Non-initial C difficile infection

0·37 (0·04–3·61)

0·45 (0·24–0·84)*

1·50 (0·06–40·63)

1·80 (0·86–3·75)

0·76 (0·18–3·23)

1·74 (0·90–3·37)

NA

Aged at least 65 years

0·79 (0·22–2·77)

0·54 (0·38–0·77)*

NA

1·61 (1·00–2·58)

1·01 (0·39–2·60)

2·90 (1·91–4·41)*

NA

Younger than 65 years

0·26 (0·08–0·80)*

0·47 (0·31–0·71)*

NA

1·30 (0·78–2·18)

0·45 (0·20–1·02)

2·52 (1·60–3·96)*

NA

Data are effect sizes, provided as odds ratios (95% CI). *Significant interactions. NA=not available.

Table 2: Summary of subgroup analyses for sustained symptomatic cure vs vancomycin

as the best treatment for severe and mild to moderate infections with C difficile, initial C difficile infection, and patients younger than 65 years. Fidaxomicin ranked as the best treatment in non-initial C difficile infection and patients aged at least 65 years. Metronidazole was inferior to fidaxomicin in all subgroups. The full subgroup analyses and rankograms are listed in the appendix.

Discussion To our knowledge, this study provides the most up-todate and comprehensive synthesis of evidence for pharmacological treatment of C difficile. In addition to

published trials, our network meta-analysis also included results from three unpublished trials that were not included in previous pairwise meta-analyses. In the final selection stage, we excluded three high-quality RCTs47,48 that investigated the effects of standard antibiotic therapy and monoclonal antibodies against C difficile toxins versus standard antibiotic therapy for achieving a primary cure and preventing the recurrence of infections with C difficile. In these trials, participants were randomly assigned to receive either monoclonal antibody or placebo, but vancomycin, metronidazole, or fidaxomicin therapy were administered on the basis of clinical

www.thelancet.com/infection Published online July 16, 2018 http://dx.doi.org/10.1016/S1473-3099(18)30285-8

7

Articles

assessment rather than random assignment. These groups are therefore not comparable to the studies included in our meta-analysis. In our network meta-analysis, teicoplanin ranked as the best treatment on the basis of P score, ridinilazole ranked second, and fidaxomicin ranked third. However, the treatment effect estimates for teicoplanin (GRADE, very low; appendix) were only based on two small RCTs29,43 that comprised 55 individuals, with a high risk of bias, and were done in 1992 and 1996. The 95% CI of the effect of teicoplanin is wide, reflecting the relatively small number of participants contributing to the network analysis, so the results should be interpreted with caution. These RCTs29,43 used intravenous teicoplanin solution, administered orally. Since 2013, oral teicoplanin in a liquid form has been licensed for use for C difficile infection in Europe, but not in the USA.49 Oral teicoplanin and vancomycin had been investigated in an earlier cohort study by de Lalla and colleagues50 in 1989. Both antibiotics showed excellent clinical response rates (100%), but the relapse rate was 13% (in patients receiving vancomycin), versus 0% in patients receiving teicoplanin. Ridinilazole (GRADE, moderate), a C difficile-specific antibiotic, has only been studied in two RCTs32,46 that comprised 64 patients. A phase 3 trial is expected to commence in 2018. Ridinilazole did not show a high primary cure rate, but had the lowest chance of recurrence among all drugs investigated. Fidaxomicin (GRADE, high) was investigated in six RCTs8,13,26,27,44,46 that comprised almost 900 patients and has the strongest evidence base to support its use. Fidaxomicin is significantly better than vancomycin, metronidazole, bacitracin, and tolevamer in achieving a sustained cure. On the basis of our results, tolevamer and bacitracin cannot be recommended for treatment of infections with C difficile. Surotomycin and LFF571, two newly developed drugs, did not show any superiority over vancomycin. Only phase 2 trial34 results for cadazolid are fully available. However, a press release51 indicates that cadazolid did not meet its primary endpoint compared with vanco­ mycin in one of two large international phase 3 trials (NCT01987895, NCT01983683) than comprised more than 1200 patients. Since 2014, guidelines52 from the European Society of Clinical Microbiology and Infectious Diseases have recommended metronidazole as the first-line treatment for initial, mild to moderate infections with C difficile. In guidelines53 from 2018, vancomycin or fidaxomicin have been recom­mended as first-line treatment options for infections with C difficile, and metronidazole is only recommended for an initial, mild to moderate C difficile infection in settings where access to vancomycin or fidaxomicin is limited.54 In our network meta-analysis, metronidazole ranked only 11th of 13 treatments in achieving a sustained symptomatic cure, was significantly inferior to five other drugs, and was inferior to fidaxomicin 8

in all subgroup analyses. Previous reports55,56 showed high faecal metronidazole con­ centrations after its intravenous administration and proposed its use when oral administration is not possible. Results of this network meta-analysis do not support the use of metronidazole as first-line C difficile therapy in oral form, and intravenous form is equally unlikely to be effective. For non-initial C difficile infection, guidelines52 from the European Society of Clinical Microbiology and Infectious Diseases recommend vancomycin or fidaxo­micin. In our analysis, fidaxomicin had a significantly higher sustained cure rate than did vancomycin in this patient group and might be considered to be a better first-line treatment. Furthermore, an RCT by Guery and colleagues26 from 2017 compared an extended duration of fidaxomicin treatment with conventional vancomycin treatment, and showed a high sustained symptomatic cure rate due to a significantly reduced C difficile recurrence rate compared with vancomycin (in seven of 131 patients achieving a primary cure vs 30 of 136 patients). The recurrence in a subgroup of patients with NAP1/BI/027 strain of C difficile did not differ between groups treated with fidaxomicin and those treated with vancomycin in a phase 3 trial.8 However, this trial was not powered to determine the effectiveness of fidaxomicin against various C difficile strains. Use of fidaxomicin as a firstline treatment for infections with C difficile is partly supported by economic evaluations, which suggest that this treatment is more cost-effective than either vancomycin or metronidazole.57 The overall consistency of the network meta-analysis for sustained symptomatic cure was good; none of the loops showed significant heterogeneity. Nevertheless, there are several limitations to this study. First, we included all RCTs, even those without sufficient blinding. Teicoplanin, which ranked as the best treatment in the overall network meta-analysis, was lost from the network meta-analysis in a sensitivity analysis when non-blinded trials were excluded. Second, industry sponsored most trials. Exclusion of these trials would have left almost no trials to compare, and this sensitivity analysis could not be done. Third, no unified C difficile infection severity assessment system was used among the RCTs. This limitation means that subgroup assessment of mild to moderate versus severe infections with C difficile was less reliable. Finally, we included all treatments that were investigated as monotherapy against infections with C difficile, even though some of them are no longer in clinical development for C difficile treatment or their use is limited by licensing barriers. Namely, teicoplanin is not licensed for treatment of C difficile in USA, Merck has discontinued the development of surotomycin after its international phase 3 trial, and ridinilazole is yet to undergo a phase 3 trial. However, inclusion of data from these trials allows us to obtain more accurate estimates of treatment effect for the other drugs in the network meta-analysis. Given its promise in small low quality

www.thelancet.com/infection Published online July 16, 2018 http://dx.doi.org/10.1016/S1473-3099(18)30285-8

Articles

RCTs, oral teicoplanin should be investigated in a large, well designed RCT to establish its sustained symptomatic cure effect more accurately. The findings of this network meta-analysis suggest that, of the currently approved treatments, fidaxomicin has the strongest evidence for being the most effective treatment in providing a long-term cure against C difficile. Apart from affordability, there is little evidence to support use of metronidazole as a first-line treatment against infections with C difficile. Early data for ridinilazole suggest that this treatment could potentially become a new, efficacious treatment against infections with C difficile, but phase 3 trials of this drug are still in progress. Contributors TB wrote the study protocol, did the searches, study selection, data extraction, statistical analyses, and wrote the initial draft of and revised the manuscript. NEB wrote the study protocol, did the study selection, data extraction, contributed to statistical analysis, and wrote and revised the final manuscript. MHW contributed to data analysis and interpretation and writing and revisions of the final manuscript. VS developed the study, was the arbiter for the study searches and data extraction, and contributed to the statistical analysis and writing and revisions of the final manuscript. Declaration of interests MHW reports grants from Actelion, Cubist Pharmaceuticals, Astellas Pharma, Merck & Co, Sanofi Pasteur, Summit Therapeutics, Seres Therapeutics, bioMérieux, Qiagen, Pfizer, and Da Volterra; personal fees from Actelion, Cubist Pharmaceuticals, Astellas Pharma, Merck & Co, Sanofi Pasteur, Summit Therapeutics, Seres Therapeutics, bioMérieux, Qiagen, Pfizer, Valneva, Ferring, Synthetic Biologics, and Alere; a grant to his department from Alere; study initiation and sponsorship by Astellas Pharma; and non-financial support from Astellas Pharma during the conduct of the study. MHW also reports grants from Tetraphase Pharmaceuticals, Surface Skins, Abbott Laboratories, European Tissue Symposium, and Paratek Pharmaceuticals; and personal fees from AstraZeneca, Nabriva Therapeutics, Pfizer, Roche, The Medicines Company, Abbott Laboratories, Basilea Pharmaceutica, European Tissue Symposium, Bayer, Allergan, Menarini, Motif Biosciences, Paratek Pharmaceuticals, AiCuris, Antabio, Spero Therapeutics, Tetraphase Pharmaceuticals, and Surface Skins outside the submitted work. TB, NEB, and VS declare no competing interests. Acknowledgments TB was the recipient of an Academic Foundation Programme fellowship from the UK National Institute for Health Research. References 1 Lawson PA, Citron DM, Tyrrell KL, Finegold SM. Reclassification of Clostridium difficile as Clostridioides difficile (Hall and O’Toole 1935) Prévot 1938. Anaerobe 2016; 40: 95–99. 2 Evans CT, Safdar N. Current trends in the epidemiology and outcomes of Clostridium difficile infection. Clin Infect Dis 2015; 60 (suppl 2): S66–71. 3 Burke KE, Lamont JT. Clostridium difficile infection: a worldwide disease. Gut Liver 2014; 8: 1–6. 4 Lessa FC, Mu Y, Bamberg WM, et al. Burden of Clostridium difficile infection in the United States. N Engl J Med 2015; 372: 825–34. 5 Desai K, Gupta SB, Dubberke ER, Prabhu VS, Browne C, Mast TC. Epidemiological and economic burden of Clostridium difficile in the United States: estimates from a modeling approach. BMC Infect Dis 2016; 16: 303. 6 Dubberke ER, Olsen MA. Burden of Clostridium difficile on the healthcare system. Clin Infect Dis 2012; 55 (suppl 2): S88–92. 7 Venugopal AA, Johnson S. Fidaxomicin: a novel macrocyclic antibiotic approved for treatment of Clostridium difficile infection. Clin Infect Dis 2012; 54: 568–74. 8 Louie TJ, Miller MA, Mullane KM, et al. Fidaxomicin versus vancomycin for Clostridium difficile infection. N Engl J Med 2011; 364: 422–31.

9

10 11

12 13

14 15

16 17 18 19 20 21 22 23 24

25 26

27

28

29

Cornely OA, Crook DW, Esposito R, et al. Fidaxomicin versus vancomycin for infection with Clostridium difficile in Europe, Canada, and the USA: a double-blind, non-inferiority, randomised controlled trial. Lancet Infect Dis 2012; 12: 281–89. Nelson RL, Suda KJ, Evans CT. Antibiotic treatment for Clostridium difficile-associated diarrhoea in adults. Cochrane Database Syst Rev 2017; 3: CD004610. Li R, Lu L, Lin Y, Wang M, Liu X. Efficacy and safety of metronidazole monotherapy versus vancomycin monotherapy or combination therapy in patients with Clostridium difficile infection: a systematic review and meta-analysis. PloS One 2015; 10: e0137252. Di X, Bai N, Zhang X, et al. A meta-analysis of metronidazole and vancomycin for the treatment of Clostridium difficile infection, stratified by disease severity. Braz J Infect Dis 2015; 19: 339–49. Cornely OA, Nathwani D, Ivanescu C, Odufowora-Sita O, Retsa P, Odeyemi IA. Clinical efficacy of fidaxomicin compared with vancomycin and metronidazole in Clostridium difficile infections: a meta-analysis and indirect treatment comparison. J Antimicrob Chemother 2014; 69: 2892–900. Crook DW, Walker AS, Kean Y, et al. Fidaxomicin versus vancomycin for Clostridium difficile infection: meta-analysis of pivotal randomized controlled trials. Clin Infect Dis 2012; 55 (suppl 2): S93–103. Hutton B, Salanti G, Caldwell DM, et al. The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations. Ann Intern Med 2015; 162: 777–84. Higgins JPT, Green S. cochrane handbook for systematic reviews of interventions version 5.1.0. March, 2011. http://handbook-5-1. cochrane.org/ (accessed Oct 19, 2018). Rouse B, Chaimani A, Li T. Network meta-analysis: an introduction for clinicians. Intern Emerg Med 2017; 12: 103–11. Rücker G. Network meta-analysis, electrical networks and graph theory. Res Synth Methods 2012; 3: 312–24. Rücker G, Schwarzer G. Ranking treatments in frequentist network meta-analysis works without resampling methods. BMC Med Res Methodol 2015; 15: 58. Brown S, Hutton B, Clifford T, et al. A Microsoft-Excel-based tool for running and critically appraising network meta-analyses— an overview and application of NetMetaXL. Syst Rev 2014; 3: 110. Lessa FC, Gould CV, McDonald LC. Current status of Clostridium difficile infection epidemiology. Clin Infect Dis 2012; 55 (suppl 2): S65–70. Krahn U, Binder H, König J. A graphical tool for locating inconsistency in network meta-analyses. BMC Med Res Methodol 2013; 13: 35. Dias S, Welton NJ, Caldwell DM, Ades AE. Checking consistency in mixed treatment comparison meta-analysis. Stat Med 2010; 29: 932–44. Higgins JP, Jackson D, Barrett JK, Lu G, Ades AE, White IR. Consistency and inconsistency in network meta-analysis: concepts and models for multi-arm studies. Res Synth Methods 2012; 3: 98–110. Jackson D, White IR, Riley RD. Quantifying the impact of between-study heterogeneity in multivariate meta-analyses. Stat Med 2012; 31: 3805–20. Guery B, Menichetti F, Anttila VJ, et al. Extended-pulsed fidaxomicin versus vancomycin for Clostridium difficile infection in patients 60 years and older (EXTEND): a randomised, controlled, open-label, phase 3b/4 trial. Lancet Infect Dis 2017; 18: 296–307. Mikamo H, Tateda K, Yanagihara K, et al. Efficacy and safety of fidaxomicin for the treatment of Clostridioides (Clostridium) difficile infection in a randomized, double-blind, comparative phase III study in Japan. J Infect Chemother 2018; published online June 19. DOI:10.1016/j.jiac.2018.05.010. Zar FA, Bakkanagari SR, Moorthi KM, Davis MB. A comparison of vancomycin and metronidazole for the treatment of Clostridium difficile-associated diarrhea, stratified by disease severity. Clin Infect Dis 2007; 45: 302–07. Wenisch C, Parschalk B, Hasenhündl M, Hirschl AM, Graninger W. Comparison of vancomycin, teicoplanin, metronidazole, and fusidic acid for the treatment of Clostridium difficile-associated diarrhea. Clin Infect Dis 1996; 22: 813–18.

www.thelancet.com/infection Published online July 16, 2018 http://dx.doi.org/10.1016/S1473-3099(18)30285-8

9

Articles

30 Wullt M, Odenholt I. A double-blind randomized controlled trial of fusidic acid and metronidazole for treatment of an initial episode of Clostridium difficile-associated diarrhoea. J Antimicrob Chemother 2004; 54: 211–16. 31 Young GP, Ward PB, Bayley N, et al. Antibiotic-associated colitis due to Clostridium difficile: double-blind comparison of vancomycin with bacitracin. Gastroenterology 1985; 89: 1038–45. 32 Vickers RJ, Tillotson GS, Nathan R, et al. Efficacy and safety of ridinilazole compared with vancomycin for the treatment of Clostridium difficile infection: a phase 2, randomised, double-blind, active-controlled, non-inferiority study. Lancet Infect Dis 2017; 17: 735–44. 33 Mullane K, Lee C, Bressler A, et al. Multicenter, randomized clinical trial to compare the safety and efficacy of LFF571 and vancomycin for Clostridium difficile infections. Antimicrob Agents Chemother 2015; 59: 1435–40. 34 Louie T, Nord CE, Talbot GH, et al. Multicenter, double-blind, randomized, phase 2 study evaluating the novel antibiotic cadazolid in patients with Clostridium difficile infection. Antimicrob Agents Chemother 2015; 59: 6266–73. 35 Musher DM, Logan N, Hamill RJ, et al. Nitazoxanide for the treatment of Clostridium difficile colitis. Clin Infect Dis 2006; 43: 421–27. 36 Musher DM, Logan N, Bressler AM, Johnson DP, Rossignol JF. Nitazoxanide versus vancomycin in Clostridium difficile infection: a randomized, double-blind study. Clin Infect Dis 2009; 48: e41–46. 37 Dudley MN, McLaughlin JC, Carrington G, Frick J, Nightingale CH, Quintiliani R. Oral bacitracin vs vancomycin therapy for Clostridium difficile-induced diarrhea. A randomized double-blind trial. Arch Intern Med 1986; 146: 1101–04. 38 Boix V, Fedorak RN, Mullane KM, et al. Primary outcomes from a phase 3, randomized, double-blind, active-controlled trial of surotomycin in subjects with Clostridium difficile infection. Open Forum Infect Dis 2017; 4: ofw275. 39 Teasley DG, Gerding DN, Olson MM, et al. Prospective randomised trial of metronidazole versus vancomycin for Clostridium-difficile-associated diarrhoea and colitis. Lancet 1983; 2: 1043–46. 40 Lee CH, Patino H, Stevens C, et al. Surotomycin versus vancomycin for Clostridium difficile infection: phase 2, randomized, controlled, double-blind, non-inferiority, multicentre trial. J Antimicrob Chemother 2016; 71: 2964–71. 41 Louie TJ, Peppe J, Watt CK, et al. Tolevamer, a novel nonantibiotic polymer, compared with vancomycin in the treatment of mild to moderately severe Clostridium difficile-associated diarrhea. Clin Infect Dis 2006; 43: 411–20. 42 Johnson S, Louie TJ, Gerding DN, et al. Vancomycin, metronidazole, or tolevamer for Clostridium difficile infection: results from two multinational, randomized, controlled trials. Clin Infect Dis 2014; 59: 345–54. 43 de Lalla F, Nicolin R, Rinaldi E, et al. Prospective study of oral teicoplanin versus oral vancomycin for therapy of pseudomembranous colitis and Clostridium difficile-associated diarrhea. Antimicrob Agents Chemother 1992; 36: 2192–96.

10

44 Thabit AK, Alam MJ, Khaleduzzaman M, Garey KW, Nicolau DP. A pilot study to assess bacterial and toxin reduction in patients with Clostridium difficile infection given fidaxomicin or vancomycin. Ann Clin Microbiol Antimicrob 2016; 15: 22. 45 Pardi DS, Brennan R, Spinnell M, et al. The efficacy and safety of rifaximin vs. vancomycin in the treatment of C difficile infection: a randomized double-blind active comparator trial. Gastroenterology 2012; 142 (suppl 1): S599. 46 Mitra S, Chilton C, Freeman J, et al. Preservation of gut microbiome following ridinilazole vs. fidaxomicin treatment of Clostridium difficile infection. Open Forum Infectious Diseases 2017; 4 (suppl 1): S526–27. 47 Wilcox MH, Gerding DN, Poxton IR, et al. Bezlotoxumab for prevention of recurrent Clostridium difficile infection. N Engl J Med 2017; 376: 305–17. 48 Lowy I, Molrine DC, Leav BA, et al. Treatment with monoclonal antibodies against Clostridium difficile toxins. N Engl J Med 2010; 362: 197–205. 49 European Medicines Agency. Targocid and associated names. Jan 1, 2014. http://www.ema.europa.eu/ema/index.jsp?curl=pages/ medicines/human/referrals/Targocid_and_associated_names/ human_referral_000341.jsp&mid=WC0b01ac05805c516f (accessed Oct 20, 2017). 50 de Lalla F, Privitera G, Rinaldi E, Ortisi G, Santoro D, Rizzardini G. Treatment of Clostridium difficile-associated disease with teicoplanin. Antimicrob Agents Chemother 1989; 33: 1125–27. 51 Actellion. Actelion provides an update on the Phase III IMPACT program with cadazolid in CDAD. https://www1.actelion.com/ investors/news-archive?newsId=2111437 (accessed Feb 18, 2018). 52 Debast SB, Bauer MP, Kuijper EJ. European Society of Clinical Microbiology and Infectious Diseases: update of the treatment guidance document for Clostridium difficile infection. Clin Microbiol Infect 2014; 20 (suppl 2): 1–26. 53 McDonald LC, Gerding DN, Johnson S, et al. Clinical practice guidelines for Clostridium difficile infection in adults and children: 2017 update by the Infectious Diseases Society of America (IDSA) and Society for Healthcare Epidemiology of America (SHEA). Clin Infect Dis 2018; 66: 987–94. 54 Shane AL, Mody RK, Crump JA, et al. 2017 Infectious Diseases Society of America clinical practice guidelines for the diagnosis and management of infectious diarrhea. Clin Infect Dis 2017; 65: 1963–73. 55 Bolton RP, Culshaw MA. Faecal metronidazole concentrations during oral and intravenous therapy for antibiotic associated colitis due to Clostridium difficile. Gut 1986; 27: 1169–72. 56 Friedenberg F, Fernandez A, Kaul V, Niami P, Levine GM. Intravenous metronidazole for the treatment of Clostridium difficile colitis. Dis Colon Rectum 2001; 44: 1176–80. 57 Burton HE, Mitchell SA, Watt M. A systematic literature review of economic evaluations of antibiotic treatments for Clostridium difficile infection. Pharmacoeconomics 2017; 35: 1123–40.

www.thelancet.com/infection Published online July 16, 2018 http://dx.doi.org/10.1016/S1473-3099(18)30285-8