Clostridium difficile–associated diarrhea (CDAD) is a unique hospital infection that occurs almost entirely in patients who have received previous antimicrobial treatment. Although not conclusively proven, the normal gastrointestinal flora is presumably disrupted by antimicrobials, which, in turn, enables ingested spores of toxigenic C. difficile to colonize the colon, produce toxins, and cause CDAD. Epidemiologic evidence suggests that specific antimicrobials and classes of antimicrobials are not created equal in terms of CDAD risk [1]. Two risks for CDAD are of interest because of how they relate to antimicrobial use: the relative risk of CDAD associated with use of a specific antimicrobial, and the attributable risk of CDAD in a particular population. Attributable risk incorporates both the relative risk of an antimicrobial and the frequency with which that drug is used in the population. Clindamycin was the highest-risk agent for CDAD in the 1970s, but its use has decreased in US and European hospitals, with a resultant reduction in attributable risk of antibiotic-associated diarrhea and CDAD [2]. By the late 1980s and through the 1990s, cephalosporin antimicrobials, particularly those of the second and third generation, such as cefuroxime, cefotaxime, ceftazidime, and ceftriaxone, had become the agents with the highest relative risk and highest attributable risk of CDAD because of their frequent use in hospitals [1]. The risk of antimicrobial-associated CDAD is increased if C. difficile is resistant to the antimicrobial [3]. For clindamycin, C. difficile resistance is variable, and risk of CDAD associated with a clindamycin-resistant organism is increased in patients receiving clindamycin [3]. For the third-generation cephalosporins, C. difficile resistance is universal, and presumably any toxigenic C. difficile organism is capable of causing CDAD during cephalosporin administration. The good news has been that CDAD rates have markedly decreased when the use of clindamycin or third-generation cephalosporins has been reduced in hospitals where these agents were associated with high CDAD rates [3, 4]. More recently, the saga of CDAD risk and antimicrobial use has been extended to the fluoroquinolone class, as shown by Gaynes et al. [5] in this issue of Clinical Infectious Diseases and by others [6–8]. Fluoroquinolones had been considered relatively low-risk agents for CDAD on the basis of a number of observations with regard to ciprofloxacin that included poor anaerobic activity, failure of fecal emulsions treated with ciprofloxacin to support growth of C. difficile in vitro (compared with clindamycin-treated emulsions), and lack of observed CDAD or C. difficile colonization in a population of 213 patients treated with ciprofloxacin monotherapy [9]. Ciprofloxacin has been used in the United States since 1988, but it was not until 2001 that a clinical case-control study showed the association between fluoroquinolone use and risk of CDAD [6]. The study was conducted in 1998 and used 2 different sets of control subjects (the first was matched for age, sex, and admission date, and the second was matched for duration of hospital stay). Multivariable analysis implicated use of ciprofloxacin as a risk factor for CDAD for the 2 control groups (OR, 5.5 [95% CI, 1.2–24.8] and 9.5 [95% CI, 1.01–88.4], respectively) [6]. A very large matched case-control study of 199 patients conducted in Pittsburgh during 2000–2001 showed a significant association between CDAD and levofloxacin use (OR, 2.2; 95% CI, 1.3–3.7), as well as significant associations between CDAD and use of clindamycin, ceftriaxone, and cefepime [7]. Of interest, in this study, more patients (n = 118) received levofloxacin than received all of the other 3 high-risk antibiotics combined (n = 98), suggesting a high attributable CDAD risk for levofloxacin [7]. The authors collected C. difficile isolates and found high-level resistance to levofloxacin, ofloxacin, moxifloxacin, gatifloxacin, and ciprofloxacin in 90% of the isolates tested [7]. Restriction endonuclease typing of isolates revealed that 51% of the isolates were of 3 major types, and 2 of these types differed by only a single band. One additional case-control study of CDAD, done in 2001, involved inpatients in 4 Veterans Affairs Medical Centers [8]. Thirty patients with CDAD and 60 control subjects were compared in a multivariable analysis, and the only significant CDAD risk factor identified was fluoroquinolone use (OR, 12.7; 95% CI, 2.6–61.6) [8]. The fluoroquinolones used were levofloxacin (60% of subjects), ciprofloxacin (15%), and gatifloxacin (15%). The study by Gaynes et al. [5] is particularly noteworthy for several reasons. Foremost is the fact that it demonstrates a difference in CDAD risk during periods when 2 different but very similar fluoroquinolones (levofloxacin and gatifloxacin) were in use, and it demonstrates that the high risk of CDAD observed with gatifloxacin can be reduced by removing the drug from use and replacing it with a similar agent from the same class. Gaynes et al. [5] were particularly successful in implementing the process of changing from one fluoroquinolone to the other, achieving 98%–99% compliance. However, the most important finding of Gaynes et al. [5] is that, after institution of the antimicrobial use change, there followed a significant reduction in CDAD rates [5]. However, CDAD rates after use of each of the fluoroquinolones were extraordinarily high (34% for gatifloxacin and 17% for levofloxacin). Furthermore, the attributable risk associated with gatifloxacin use was very high: 14 [67%] of 21 patients who had CDAD received gatifloxacin, compared with 5 (24%) of 21 patients who received clindamycin, the second-most frequently associated antimicrobial. In addition, risk of CDAD in the long-term care facility increased with the duration of gatifloxacin therapy. Rates of CDAD per 1000 patient-days were 3–4 times higher in the long-term care facility and in the acute care hospital during the period of gatifloxacin use (as well as the first period of levofloxacin use in the hospital), compared with 10-year historical CDAD rates for the acute care section of another Veterans Affairs hospital for the period of 1982–1991, before most fluoroquinolones were available [10]. What is driving this apparent increase in CDAD after use of newer fluoroquinolones? The authors speculate that the enhanced anaerobic spectrum of newer fluoroquinolones, such as gatifloxacin, may have a more disruptive effect on the fecal flora [5]. Ironically, these agents are usually much more active against C. difficile than are older fluoroquinolones, such as ciprofloxacin [11]. Gaynes et al. [5] found that 42 of 45 C. difficile isolates were resistant to levofloxacin, moxifloxacin, and gatifloxacin (MICs, >32 mg/L), which is remarkably similar to the finding that 90% of isolates had high-level resistance in the Pittsburgh outbreak associated with levofloxacin use [8]. Gaynes et al. [5] also found evidence of clonality; 25 of 45 C. difficile isolates of type A, and in the Pittsburgh study, 51% of C. difficile isolates were of 3 types [5, 8]. Neither study has shown that the clonal isolates were more likely to be resistant to the fluoroquinolones, but this was shown by Wilcox et al. [11] for the predominant C. difficile clone in an epidemic in the United Kingdom. Obviously, there is need to study the epidemiology of these CDAD outbreaks more carefully and to collect C. difficile isolates for typing and antimicrobial susceptibility testing. I predict that, when this is done, it will be confirmed that clones of C. difficile that are resistant to fluoroquinolones are responsible for an increasing number of CDAD outbreaks associated with the widespread hospital use of newer fluoroquinolones. I agree that the anaerobic activity of newer fluoroquinolones is likely to be more disruptive of normal flora, but I believe that the additional factors of acquired fluoroquinolone resistance in C. difficile and proliferation of these resistant clones in specific hospitals are the critical events leading to these outbreaks. As has been observed with many previous instances of infections with antimicrobial-resistant organisms and CDAD outbreaks, the good news, as demonstrated by Gaynes et al. [5], is that removal of the offending antimicrobial can result in a marked reduction in the rate of CDAD. Whether substitution of so closely related a fluoroquinolone as levofloxacin will be successful elsewhere is unclear remains to be seen.
No takes yet. Share an insight, caveat, or question.
Dale N. Gerding (2004) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: