Efficacy and Safety of Metronidazole Monotherapy versus Vancomycin Monotherapy or Combination Therapy in Patients with Clostridium difficile Infection: A Systematic Review and Meta-Analysis

Background Clostridium difficile infection (CDI) has become a global epidemiological problem for both hospitalized patients and outpatients. The most commonly used drugs to treat CDI are metronidazole and vancomycin. The aim of this study was to compare the efficacy and safety of metronidazole monotherapy with vancomycin monotherapy and combination therapy in CDI patients. Methods A comprehensive search without publication status or other restrictions was conducted. Studies comparing metronidazole monotherapy with vancomycin monotherapy or combination therapy in patients with CDI were considered eligible. Meta-analysis was performed using the Mantel-Haenszel fixed-effects model, and odds ratios (ORs) with 95% confidence intervals (95% CIs) were calculated and reported. Results Of the 1910 records identified, seventeen studies from thirteen articles (n = 2501 patients) were included. No statistically significant difference in the rate of clinical cure was found between metronidazole and vancomycin for mild CDI (OR = 0.67, 95% CI (0.45, 1.00), p = 0.05) or between either monotherapy and combination therapy for CDI (OR = 1.07, 95% CI (0.58, 1.96), p = 0.83); however, the rate of clinical cure was lower for metronidazole than for vancomycin for severe CDI (OR = 0.46, 95% CI (0.26, 0.80), p = 0.006). No statistically significant difference in the rate of CDI recurrence was found between metronidazole and vancomycin for mild CDI (OR = 0.99, 95% CI (0.40, 2.45), p = 0.98) or severe CDI (OR = 0.98, 95% CI (0.63, 1.53), p = 0.94) or between either monotherapy and combination therapy for CDI (OR = 0.91, 95% CI (0.66, 1.26), p = 0.56). In addition, there was no significant difference in the rate of adverse events (AEs) between metronidazole and vancomycin (OR = 1.18, 95% CI (0.80, 1.74), p = 0.41). In contrast, the rate of AEs was significantly lower for either monotherapy than for combination therapy (OR = 0.30, 95% CI (0.17, 0.51), p<0.0001). Conclusions Metronidazole and vancomycin are equally effective for the treatment of mild CDI, but vancomycin is superior for the treatment of severe CDI. Combination therapy is not superior to monotherapy because it appears to be associated with an increase in the rate of AEs.


Introduction
Clostridium difficile is an anaerobic, gram-positive, spore-forming bacillus [1]. The incidence and severity of C. difficile infection (CDI) appear to be on the rise [2], and CDI has become an epidemiological problem for both hospitalized patients and outpatients worldwide [2].
Oral metronidazole is recommended for patients undergoing initial therapy and for those with mild infections, whereas oral vancomycin is recommended for seriously ill patients. Furthermore, the combination of oral vancomycin and intravenous metronidazole is suggested for the treatment of severe or complicated infections [3,4]. As risk of the development of vancomycin-resistant enterococci (VRE) associated with both metronidazole and vancomycin treatment has emerged [5], studies examining fidaxomicin [6,7], rifaximin [8,9], rifampin [10], nitazoxanide [11,12], tolevamer [13], and teicoplanin [14] therapies have been performed. Furthermore, in patients with recurrent CDI, infusion of donor feces has resulted in improved treatment outcomes. In particular, patients with multiple relapses of CDI have benefited from this approach [15]. Unfortunately, prospective clinical trials of most of the above-mentioned antibiotics evaluating their applicability for general use have not yet been performed [4], and several questions regarding fecal transplantation remain unanswered; for example, the optimal protocol for donor feces infusion is unknown [15]. Therefore, the treatment of CDI has relied primarily on metronidazole and vancomycin [16].
Although the Healthcare Epidemiology of America (SHEA)/Infectious Diseases Society of America (IDSA) guidelines recommend the use of oral vancomycin for severe CDI, this recommendation is based on the results of few clinical trials [17,18]. In addition, despite the frequent use of combination therapy in clinical settings, it is unclear whether combination therapy is more effective than monotherapy because, as emphasized by the European Society of Clinical Microbiology and Infectious Diseases (ESCMID) [4], few randomized controlled trials (RCTs) have been performed to compare them. Moreover, the many associated adverse events (AEs) [16] necessitate safety analyses of these two drugs. To address this knowledge gap, we performed a systematic review and meta-analysis to compare the efficacy and safety of metronidazole monotherapy with vancomycin monotherapy and combination therapy for the treatment of patients with CDI.

Methods
Our study protocol and analysis were planned in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [19], and no ethical approval was needed.

Search strategy
A systematic literature search of the following electronic databases was conducted to identify relevant literature published in English or Chinese before November 2014: PubMed, Embase, Web of Science, and Cochrane Library. Three Chinese language databases-China National Knowledge Infrastructure (CNKI; available at www.cnki.net), Chinese Scientific Journals Database (VIP; available at www.cqvip.com), and WANFANG DATA (available at www. wanfangdata.com.cn)-were also searched. The search strategy included the following medical subject headings and key words: "pseudomembranous enterocolitis," "antibiotic-associated diarrhea," "antibiotic-associated colitis," "Clostridium difficile," "metronidazole," and "vancomycin." These search terms were combined using Boolean logic as follows: both the medical subject headings and the terms related to the patient population of interest (Clostridium difficile, pseudomembranous colitis, antibiotic-associated diarrhea, and antibiotic-associated colitis) were combined with those describing interventions (metronidazole OR vancomycin). The complete search strategy used to search each database is described in S1 Table. Furthermore, the references in the initially identified articles, including relevant reviews, were manually searched and reviewed. Academic dissertations and abstracts presented at scientific conferences were also searched using ProQuest Dissertation & Theses Database and The Conference and Academic Dissertation Database of CNKI, respectively, to ensure that no relevant study was missed (up to May 2015).

Study selection
Two reviewers (RL and LCL) assessed all potentially relevant studies and reached a consensus on all items. They then independently searched the literature and examined relevant studies to obtain data on the rates of clinical cure, CDI recurrence and AEs. In the case of disagreement between the two reviewers, the senior coauthor (XL) was consulted, and the disagreement was resolved by consensus. Studies were included if they met the following criteria: (1) contained original data, (2) contained data regarding the use of metronidazole or vancomycin for the treatment of CDI, (3) contained data regarding clinical therapeutic outcomes or AEs, and (4) reported on RCTs or case-control studies. Studies were excluded if they (1) were unrelated to CDI, (2) were duplicate reports, (3) were not written in English or Chinese, or (4) were not case-control studies.

Data extraction
Two reviewers (RL and LCL) independently abstracted data from each eligible study. Each investigator was blinded to the other investigator's extracted data. A standardized form was used to record the following information: (1) first author; (2) year of publication; (3) disease severity (although there were four classifications according to disease severity, including mild CDI, severe CDI, complicated CDI, and recurrent CDI [3,4], none of the included studies evaluated complicated CDI, and fecal microbiota transplantation has been performed for recurrent CDI. Additionally, one included study referred to "mild-moderate" because the treatment of mild CDI and moderate CDI is identical [20], and we categorized this classification of CDI into the mild CDI group; thus, this systematic review and meta-analysis was primarily focused on the treatment of mild or severe CDI); (4) mean age; (5) male (%); (6) follow-up duration; (7) number of enrolled patients; (8) drug regimen; (9) rate of clinical cure; (10) rate of CDI recurrence; (11) AEs related to the study medications; (12) case definitions; (13) quality of the evidence; and (14) overall risk of bias. The extracted data (and specifically, the rates of clinical cure and CDI recurrence) were analyzed based on the intention-to-treat population.

Quality appraisal
The included studies were independently appraised for methodological quality by two authors (RL and LCL), without blinding to the source journal or authorship. Discrepancies were resolved by discussion or consultation with the third reviewer (XL) if required. The quality of each included study was evaluated according to the modified Jadad score [21]. The overall risk of bias of each included study was also evaluated [22]. Potential publication bias was assessed by visual inspection of asymmetry in Begg's funnel plots, and Egger's test was then used to provide statistical evidence of funnel plot symmetry (p<0.05 indicating bias and p>0.05 not indicating bias) [23,24]. We also performed sensitivity analyses on subgroups from the studies according to explicit or inexplicit classification of severity and follow-up duration to explore underlying sources of heterogeneity. All of these analyses, except for publication bias analysis (which was performed using STATA software: version 12.0, StataCorp, College Station, TX, USA), were performed using Review Manager software (RevMan, version 5.1, Oxford, UK; The Cochrane Collaboration, 2008).

Outcomes analyzed
Regarding the outcome measures used to assess efficacy, in this meta-analysis, the rate of clinical cure was used as the primary outcome measure, and the rate of CDI recurrence was used as the secondary outcome measure. We present the following two comparisons based on the rates of clinical cure and CDI recurrence: a comparison between metronidazole and vancomycin monotherapy for the treatment of patients with mild or severe CDI and a comparison of vancomycin or metronidazole monotherapy with combined vancomycin and metronidazole therapy or either drug combined with another antibiotic for the treatment of patients with CDI. For the latter comparison, we performed subgroup analysis according to the therapeutic intervention. AEs were used as the primary safety outcome measure in this meta-analysis. Here, we also present the following two comparisons based on AEs: metronidazole vs. vancomycin and monotherapy vs. combination therapy. We performed subgroup analysis according to AEs for both comparisons.

Data analysis and statistical methods
Statistical analysis was accomplished using Review Manager version 5.1 software. Both I 2 and Q statistics were considered. In our analyses, we used the I 2 statistic to quantitatively describe heterogeneity across studies. An I 2 value of greater than 50% indicated high heterogeneity; a value ranging from 25% to 50% indicated moderate heterogeneity; and a value of less than 25% signified low heterogeneity [22]. Pooled odds ratios (ORs; calculated by adding 0.5 to each cell of the 2×2 table for the trial when one arm of the study contained no events [25]) and 95% confidence intervals (CIs; if 95% CIs did not include the null value, the results were considered to be significantly different) were also used in meta-analysis. The Mantel-Haenszel fixed-effects model was used for analysis of the total groups because the I 2 values in our subgroup metaanalysis ranged from 0% to 6%.

Search results
A total of 1910 records were identified (Fig 1). After excluding duplicates and screening the titles of the studies, 635 articles were reviewed. After screening the abstracts of these potentially relevant articles, 36 were selected for full-text review based on relevance to the study topic (Fig  1). Thirteen articles containing seventeen studies (one article included three separate studies and two articles each included two separate studies) and 2501 patients were included [10,13,20,[26][27][28][29][30][31][32][33][34][35]. In addition, we attempted to contact the corresponding authors of five potentially relevant studies via e-mail. The authors of four studies did not respond to our e-mail communication [36][37][38][39], and the corresponding author of the fifth study communicated that we would require approval by their institutional review board for additional individual patient data [40]. Therefore, all five studies had to be excluded from our final set of included studies. Fig 1 shows the selection process for the studies included in the meta-analysis.

Publication bias and sensitivity analyses
The shape of Begg's funnel plot and the Egger's test results (all p>0.05) did not demonstrate any evidence of publication bias (S1 File). Considering that meta-analysis showed little heterogeneity (I 2 = 6%) between the studies for the comparison of the rate of CDI recurrence between metronidazole and vancomycin. Thus, we conducted sensitivity analyses of the subgroups from these studies according to explicit or inexplicit classifications of severity and follow-up duration to explore underlying sources of heterogeneity. These analyses showed that the primary results were not influenced by the inexplicit classification of severity or by the follow-up durations (S2 File).

Discussion
The aim of this study was to compare the efficacy and safety of metronidazole monotherapy with vancomycin monotherapy and combination therapy in CDI patients. Based on the analyses according to disease severity and the treatment methods used, this study has revealed that metronidazole and vancomycin are equally effective for the treatment of mild CDI but that vancomycin is superior for severe CDI. Additionally, these findings suggest that combination therapy is not superior to monotherapy. Regarding efficacy, the results of this meta-analysis demonstrated that the treatments were equivalent for mild disease but not for severe disease. Moreover, regarding the cost of the medications, metronidazole is less expensive than vancomycin [41], and oral vancomycin may be more likely to promote the generation and overgrowth of VRE than metronidazole [5]. As a result, metronidazole is recommended for treating mild CDI, and vancomycin is recommended for treating severe CDI. This meta-analysis also demonstrated that combination therapy might not be more effective than monotherapy. Regarding safety, although there was no statistically significant difference in the rate of AEs between metronidazole and vancomycin, there was an increase in the rate of AEs in association with combination therapy, and the combination therapy group experienced more fatal AEs, such as death and complications. Therefore, monotherapy appears to be safe and well tolerated. Nonetheless, it is worth mentioning that because of the limited number of included studies, we did not compare monotherapy with combination therapy for all four categories of disease severity. Moreover, our results may have been influenced by the medical conditions or treatment (e.g., nasogastric feeds [10]) of the patients receiving combination therapy, who may have suffered from more serious illnesses than those receiving monotherapy; therefore, these influences may have been associated with the increased risk of experiencing AEs or developing complications among those receiving combination therapy.
Our study has several limitations. First, although the quality of the majority of the studies was high or medium and although the total sample size was sufficient, the sample size of each subgroup was relatively small and was thus susceptible to false-positive or false-negative results. Second, although sensitivity analyses indicated that the primary results were not influenced by the inexplicit classification of severity or by the follow-up duration, we roughly determined disease severity based on the proportion of patients, which potentially resulted in bias. For example, if 77% of the patients had mild CDI [27], we categorized all of the patients into the mild CDI group. Moreover, the follow-up duration for CDI treatment varied greatly (from 3-12 weeks) among the different studies, which may have resulted in bias. Third, because very few articles presented research on combination therapy, we could not compare the subgroups according to disease severity between monotherapy and combination therapy; therefore, the general patient conditions may have caused prescription bias. Fourth, only studies published in English were included, which might have rendered the results vulnerable to bias related to language and ethnicity. Fifth, additional data on combination therapy using other novel antibiotics are needed.
In conclusion, these findings have important clinical implications and support the recommendations of the current CDI treatment guidelines [3,4]. However, validation of the routine use of combination therapy may require additional evidence. To compensate for the shortcomings of our study, further large-scale clinical trials and well-designed research studies are needed to identify more effective therapies for patients with CDI.
Supporting Information S1 PRISMA Checklist. PRISMA checklist for the meta-analysis.