<i>Editorial Commentary:</i>Decreased Effectiveness of Metronidazole for the Treatment of<i>Clostridium difficile</i>Infection?

Since the emergence of a new virulent strain of Clostridium difficile—characterized as toxinotype III, North American PFGE type 1 (NAP1), restriction endonuclease analysis group type BI, and PCR ribotype 027 (type 027)—multiple outbreaks have been reported in North America and Europe [1, 2]. The first reports were from Canada, where the province of Quebec was the earliest and most severely affected [3]. In the United States, at least 38 states have been affected by the new emerging strain [4]. Outbreaks in the United Kingdom and The Netherlands have been followed by a rapid spread of the strain throughout Europe, now including 18 countries [5]. A worrisome new development is outbreaks in Europe attributable to clindamycin-resistant NAP1/027 strains (MIC, >256 mg/L) that harbor the ermB gene [6]. Clindamycin has been considered to be a “protective” antibiotic for the development of C. difficile–associated disease (CDAD) attributable to NAP1/027, but resistance to this agent increases the risk of C. difficile infection in patients, and the use of this agent may be an important factor contributing to the persistence and spread of this strain [7]. Most studies of NAP1/027 report an increased severity of C. difficile infection, a high relapse rate, and significant mortality [1–3, 7, 8]. This increased virulence is thought to be associated with a deletion at nucleotide position 117 of the tcdC gene that leads to an increased or prolonged production of toxins A and B and possibly to the production of a binary toxin [1, 9]. However, these virulence factors are not unique for NAP1/027 and are also present in other PCR ribotypes, such as type 078. Interestingly, the National Reference Laboratory for C. difficile in The Netherlands reported an increase in C. difficile PCR ribotype 078, which is currently the third-most frequently encountered type in cases of human CDAD (after NAP1/027 and PCR ribotype 014) [10]. By contrast, the most frequently found strain in the United Kingdom is PCR ribotype 106, yet this strain does not contain any of the above-mentioned virulence markers. C. difficile PCR ribotype 106 has not yet been reported in other countries. These observations emphasize the importance of using continuous surveillance systems for C. difficile infection in individual countries to observe not only changes in the epidemiology of individual strains but also potential alterations in disease presentation and response to therapy. The first step in treatment of a patient with documented or suspected C. difficile infection is to discontinue treatment with the offending antimicrobial, when possible. Antiperistaltic and opiate agents should be avoided, especially in the acute infection setting. Metronidazole has long been the first-line agent in the treatment of CDAD, but a decreasing effectiveness of metronidazole in CDAD treatment has been documented in a prospective observational study in which 22% of patients treated with ⩾1.5 g of metronidazole per day had persistent symptoms after 10 days of metronidazole treatment [11]. In Quebec, the rate of suboptimal response to metronidazole more than doubled (to 25.7%) during 2003–2004, compared with the response during 1991–2002 (9.6%) [8]. Metronidazole does not achieve high levels in feces. Metronidazole therapy (400 mg every 8 h) for 9 patients resulted in mean levels (±SD) in feces of 9.3±7.5 μg/g wet weight (range, 0.8–24.2 μg/g), with higher concentrations in watery samples or semiformed samples than in formed samples [12]. The hydroxymetabolite of metronidazole, which has also activity against C. difficile, is detectable at roughly the same levels. Given the poor fecal concentrations of metronidazole achieved after a dosage of 400 mg every 8 h, it is illogical to use lower dosages (e.g., 250 mg every 6–8 h). Furthermore, even a modest increase in the MIC of metronidazole for C. difficile might result in insufficient fecal antibiotic concentrations to inhibit (vegetative) bacteria. By contrast, oral vancomycin (125–250 mg every 6 h) administration generally yields fecal levels >1000 mg/L in patients with C. difficile infection [13]. Intravenously administered vancomycin does not result in sufficient intestinal concentrations and is not a therapeutic option. The 3 most frequently recognized risk factors for severe C. difficile infection are age, peak leukocytosis, and serum creatinine level [14, 15]. Treatment algorithms have been proposed that recommend that mild-to-moderate or severe cases of C. difficile infection be treated with metronidazole (500 mg 3 times per day) or vancomycin (125–250 mg every 6 h), respectively [14]. That study defined severe C. difficile infection on the basis of a score that was based on patient age, temperature, albumin level, peripheral WBC count, endoscopic evidence of pseudomembranous colitis, or treatment in the intensive care unit. However, definitive severity markers await confirmation in prospective studies that would preferably include a weighting for patient comorbidities. Until this is available, we recommend that any of the following may indicate severe C. difficile infection: WBC count >15,000 cells/μL, acutely rising serum creatinine level (e.g., >50% increase above baseline), elevated serum lactate concentration [16], temperature >38.5°C, or evidence of severe colitis (e.g., abdominal signs or radiology). This issue of Clinical Infectious Diseases reports a 9-month prospective observational study of the clinical and microbiological responses to treatment of C. difficile infection during 10 days of treatment with metronidazole and/or vancomycin [17]. The open study comprised a modest 52 patients total, of whom ∼60% were infected with the NAP1/027 strain. The authors concluded that vancomycin-treated patients showed a better microbiological response between days 1 and 5 of therapy, as measured by eradication of C. difficile and resolution of diarrhea, compared with patients treated with metronidazole alone. Eradication of C. difficile was achieved in 53% of vancomycin-treated patients and in 30% of metronidazole-treated patients at day 5, but at the end of therapy, both groups had similar percentages of eradication (87.5%). No differences occurred in mortality or recurrence rate for 3 months after the diagnosis of C. difficile infection. The question arises whether the results of this study should lead to the recommendation to use oral vancomycin to treat all patients with C. difficile infection. The reasons for initiating treatment with vancomycin or metronidazole were not clear. We assume that the large number of patients given vancomycin reflects the fact that this antibiotic was used preferentially in cases of a previous C. difficile infection episode. The 2 patient groups differed in patient age, previous episodes of CDAD (P=.002), serum albumin concentration (P=.026), and stay in the intensive care unit within 2 weeks after C. difficile infection diagnosis (not significant). Unfortunately, the dosages and duration of treatments of metronidazole and vancomycin are not detailed; therefore, it is unclear whether equal proportions of patients received sufficient antibiotic or were compliant with therapy. Indeed, the authors acknowledge that the metronidazole treatment failure rate may have been overestimated because some metronidazole-treated patients did not receive a sufficiently long course of antibiotic before vancomycin was substituted. However, this study supports the need to assess the response of the patients at least daily and to switch treatment to vancomycin if symptoms worsen or have not been resolved by day 6–7. It also demonstrated that proton pump inhibitors may be associated with C. difficile infection treatment failures, although this issue clearly requires further study. What is the reason for the decreasing effectiveness of metronidazole? C. difficile infection caused by C. difficile ribotype 027 was associated with a significantly higher metronidazole failure rate than were cases attributable to other C. difficile types in a study of matched cases [18]. Al-Nassir et al. [17] did not show a difference in antibiotic response by C. difficile type, although the small cohort size may be relevant here. Metronidazole failure may be due to low dosage and/or poor pharmacokinetics. Ideally, Al-Nassir et al. [17] should have measured metronidazole levels in fecal samples. Notably, there is evidence that inactivation of metronidazole occurs in the presence of gut contents, possibly because of metabolism by enterococci [19, 20]. Confirmed evidence of metronidazole resistance has hitherto been rare. However, we have recently detected the emergence of reduced susceptibility to metronidazole in C. difficile strains in the United Kingdom [21]. C. difficile strains isolated from symptomatic patients in Leeds, United Kingdom, in 2005–2006 were screened for metronidazole resistance, and the results were compared with those from repeat testing of historical isolates (1995–2001). Isolates of C. difficile ribotypes 001 (87 isolates), 106 (81), and 027 (48), the most common types seen in the United Kingdom, and 10 other prevalent ribotypes (57) were examined, with initial use of a spiral gradient end-point method. No reduced susceptibility to metronidazole was observed in C. difficile ribotype 106 or 027 isolates (geometric mean MICs by spiral gradient end-point testing, 1.1 mg/L and 0.9 mg/L, respectively). By contrast, 21 (24%) of C. difficile ribotype 001 isolates had reduced susceptibility to metronidazole by spiral gradient end-point testing (geometric mean MIC, 3.5 mg/L [P<.001]). No reduced susceptibility to vancomycin was observed. Crucially, neither Etest nor Clinical and Laboratory Standards Institute methods detected the C. difficile strains with reduced susceptibility to metronidazole. By contrast, reduced susceptibility to metronidazole was be confirmed using an agar dilution method. Discrepancies between metronidazole MICs measured by agar incorporation and Etest have been reported elsewhere for C. difficile [22, 23]. Highly discriminatory DNA fingerprinting with use of multilocus variable number of tandem repeat analysis suggested the dissemination of 3 subgroups with the reduced metronidazole resistance phenotype [24]. The exact mechanism of reduced susceptibility to metronidazole remains to be determined. Al-Nassir et al. [17] used the Etest to determine the susceptibility of C. difficile isolates to metronidazole and did not find evidence of reduced susceptibility, and that could, in part, explain the poor microbiological response. The method used to measure metronidazole susceptibility is important, and increased vigilance will be required to identify reduced antibiotic susceptibility in C. difficile. Observational studies have marked limitations, and the results are sometimes contradictory. For example, we note that discrepant results have been recently reported by Pépin et al. [25], who found that vancomycin had lost its superiority over metronidazole for NAP1/027, although no information was collected concerning the dosages used. There is mounting evidence that metronidazole is inferior to vancomycin, at least for some patients with C. difficile infection (i.e., those with severe infection). C. difficile infection guidelines, soon to be issued, will include treatment recommendations that are based on severity assessment. Definitive conclusions for the optimal treatment of C. difficile infection can be obtained only through randomized controlled trials that account for differences in case severity and other potential confounding factors, including host response. We thank Prof. Paola Mastrantonio (Department of Infectious, Parasitic, and Immune-Mediated Diseases, Istituto Superiore di Sanità, Rome, Italy) for support in the laboratory tests for metronidazole susceptibility. Potential conflicts of interest. M.H.W. has received honoraria for consultancy work, financial support to attend meetings, and research funding from Astra-Zeneca, Bayer, Genzyme, Nabriva, Pfizer, Vicuron, and Wyeth. E.J.K.: no conflicts.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.