Individual members of the fluoroquinolone class demonstrate different spectra of activity and pharmacokinetic profiles. The first-generation fluoroquinolones (e.g., ciprofloxacin, ofloxacin, norfloxacin) are primarily active against gram-negative and some gram-positive organisms. The second-generation fluoroquinolone, levofloxacin, is the L-isomer of ofloxacin and demonstrates somewhat improved gram-positive activity. However, susceptibility data show levofloxacin to be less potent than ciprofloxacin against such gram-negative pathogens as Pseudomonas aeruginosa and certain enterobacteriaceae (12). The third-generation fluoroquinolones include moxifloxacin and gatifloxacin and have improved gram-positive, atypical, and anaerobic coverage compared with first- and second-generation fluoroquinolones. In particular, these newer representatives of the fluoroquinolone class manifest greater activity against Streptococcus pneumoniae, an important respiratory pathogen (12).

The relative activities of these fluoroquinolones, expressed as 90% MICs (MIC₉₀s), are shown in Table 1. Ciprofloxacin is the most active fluoroquinolone against P. aeruginosa; typical MICs of ciprofloxacin are two- to eightfold lower than those of levofloxacin or newer quinolones such as moxifloxacin and gatifloxacin (12–16). Species of enterobacteriaceae also differ in their susceptibility to the quinolones (12). Ciprofloxacin is generally twofold more active against Escherichia coli and Klebsiella pneumoniae than levofloxacin and moxifloxacin (Table 1).

Conversely, ciprofloxacin (1.0–4.0 mg/L) and levofloxacin (1.0–2.0 mg/L) are not as active against S. pneumoniae as moxifloxacin (0.06–0.25 mg/L) or gatifloxacin (0.5–1.0 mg/L) (12–16). A recent survey conducted in the United States and Canada showed ciprofloxacin MIC₉₀s of 2 mg/L against S. pneumoniae to be identical to those of levofloxacin but higher than those of the third-generation fluoroquinolone gatifloxacin (0.5 mg/L) (17). In addition to improved gram-positive activity, third-generation fluoroquinolones have improved activity against some anaerobic species compared to first- and second-generation fluoroquinolones. MIC₉₀s for Prevotella/Porphyromonas, Fusobacterium species, and Peptostreptococcus species for levofloxacin are 1.0–8.0, 1.0–8.0, and 4.0 mg/L, as compared with moxifloxacin (0.5–4.0,0.125–4.0, and 0.5 mg/L), respectively (18). Activity of newer fluoroquinolones against a variety of atypical organisms is also improved. For example, for Mycoplasma pneumoniae, MICs are 1.0 mg/Land 0.5 mg/L for ciprofloxacin and levofloxacin, respectively, and 0.125 mg/L for both moxifloxacin and gatifloxacin (19).

Fluoroquinolone MIC₉₀s will increase as resistant mutants invariably emerge, although the rate at which resistance develops largely depends on appropriate use. Patient-, institution-, and infection-specific therapeutic decisions require that antimicrobial susceptibilities be routinely tested and reported. Accurately assessing these MIC changes depends on the precision of the test used. The standard doubling dilution techniques used in most hospital microbiology laboratories may not be precise enough to identify minor susceptibility changes within a bacterial population (20). Utilizing the E-test method, which is sensitive enough to detect these subtle MIC changes (21), as follow-up for monitoring and controlling resistant strains isolated with increasing frequency in the clinical laboratory might be a practical solution, even though this approach requires greater resource and acquisition costs. Detecting and reporting these susceptibility changes are important since they can predict changes in the resistance potential of a pathogen (22). These data may be used to develop appropriate prescribing patterns to preserve antimicrobial activity.

Moreover, susceptibility data may not be accurate because surrogate methods, such as class-representative disk testing, are used in many institutions (23). Fuchs et al. (24) found that an accurate prediction of levofloxacin resistance could not be derived from use of ciprofloxacin or ofloxacin disk testing. This study showed that levofloxacin MICs were underestimated. Accordingly, the drug whose clinical use is being considered must be tested directly.

After observing three failures in patients treated with levofloxacin for pneumococcal infections, Davidson et al. conducted a survey in 2000 (25) and found that 86% of Canadian laboratories tested only nonfluoroquinolone antimicrobial agents for S. pneumoniae susceptibility. Given the growing resistance to traditional first-line agents and the increasing number of guidelines promoting quinolones as an alternative first-line choice in some patients (26–28), relevant testing should be routinely performed. Highlighting the need for fluoroquinolone susceptibility testing, Sahm et al. (29) noted a significant (p<0.005) increase in pneumococcal levofloxacin resistance in1997–1998 and 1998–1999 from 0.1% to 0.6%, although the incidence of S. pneumoniae resistance to fluoroquinolones remains low (<1%) in the United States (30,31).

Pathogenic bacteria employ a variety of strategies to persist and replicate under adverse conditions such as exposure to an antimicrobial agent. The efflux pump system is a mechanism that allows immediate survival of bacteria in the presence of an antimicrobial agent by actively expelling that agent across the cell membrane, thereby reducing the intracellular concentrations to sublethal levels. The pump’s action is dependent on the antimicrobial’s ability to bind to the bacterial efflux protein and be exported. Some fluoroquinolones, such as moxifloxacin and trovafloxacin, are not as affected by bacterial efflux mechanisms because of their bulky side-chain moiety at position 7, which hinders export (32).

Another resistance mechanism involves specific point mutations that reduce the binding of the antimicrobial agent to specific enzymatic sites by altering the target site. In this regard, fluoroquinolones bind to enzymes involved in DNA replication, including DNA gyrase and DNA topoisomerase IV. Specific mutations in the genes that code for these enzymes can result in decreased binding and activity of the fluoroquinolones (33). Different fluoroquinolones demonstrate stronger or weaker affinity to these enzyme-binding sites. First- and second-generation fluoroquinolones bind primarily to DNA gyrase or DNA topoisomerase IV, depending on the bacteria and the drug, whereas the third-generation fluoroquinolones generally bind strongly to both DNA gyrase and DNA topoisomerase IV. Thus, a single point mutation in DNA gyrase or DNA topoisomerase IV generally affects first- and second-generation fluoroquinolones to a greater extent than third-generation fluoroquinolones. Furthermore, the third-generation C-8 methoxyfluoroquinolones, moxifloxacin and gatifloxacin, appear to bind different molecular sites within these enzymes, thereby decreasing the cross-resistance between these agents and the older fluoroquinolones (34,35).

The rate at which resistance develops to an antimicrobial agent is a measure of the resistance potential of the agent and can be assessed in vitro. M’Zali et al. (36) compared the mutant selecting potential of ciprofloxacin and levofloxacin in Pseudomonas aeruginosa. In this study, clinical isolates of P. aeruginosa were repeatedly exposed to concentrations below the MICs for ciprofloxacin and levofloxacin. The fluoroquinolone-resistant strains emerged at a significantly increased rate with levofloxacin compared to ciprofloxacin (p<0.001). These findings were consistent with those of Gilbert et al. (37).

Likewise, Dalhoff et al. (38) compared the resistance selection potential of various fluoroquinolones in vitro after repeated overnight exposures to suboptimal concentrations of S. pneumoniae. In this study, the C-8-methoxyquinolones (moxifloxacin and gatifloxacin) showed a lower propensity to select resistant mutants compared with levofloxacin and ofloxacin.

Appropriate Fluoroquinolone Selection: Pharmacokinetic and Pharmacodynamic Considerations

Pharmacokinetic properties, including the concentration of drug in the serum over time (area under the curve [AUC]) and the peak serum concentration of the drug (Cmax), can be measured, and when considered in combination with in vitro activity, may be useful for predicting microbiologic success and clinical outcomes. In particular, the ratio of the Cmax to MIC or AUC to MIC (AUIC) can be predictive of drug efficacy, although which parameter is most predictive of clinical outcome is the subject of some disagreement. Generally, the higher the ratio, the better the outcome (39,40).

While fluoroquinolones are generally concentration-dependent bactericidal agents, differences in antibacterial activities exist among class members. Fluoroquinolones also differ in pharmacokinetic parameters, such as Cmax and AUC (39). These efficacy parameters, as they relate to S. pneumoniae and P. aeruginosa, for ciprofloxacin, levofloxacin, moxifloxacin, and gatifloxacin are shown in Table 2. Cmax/MIC and AUIC are highest for ciprofloxacin against P. aeruginosa; against S. pneumoniae, these values are highest with moxifloxacin.

Although AUC/MIC and Cmax/MIC ratios are useful for predicting antimicrobial efficacy, they may not be as useful for predicting the potential for drug resistance to develop. In this regard, Thomas et al. (45) suggest that AUC/MIC should exceed 100 for gram-positive and gram-negative species to prevent resistance selection.

Alternatively, Zhao et al. (46) have hypothesized that the rate at which resistance develops to a fluoroquinolone is related to its MICs and mutant prevention concentrations. Studies involving a range of bacterial species suggest that the concentration to prevent mutant emergence in the clinical setting can be derived in vitro and is 2 to 4 times higher than the MIC for most fluoroquinolones (46); however, the clinical significance of these findings has not been clearly established. Derivation of the mutant prevention concentrations is a process involving spreading a high bacterial load onto a series of agar plates in which various concentrations of antimicrobial have been incorporated. The density of 10¹⁰ CFU/mL was selected to pinpoint frequency of mutation at levels of 10^-7, 10^-8, and 10^-9, as well as to model the bacterial load at the site of infection. The inoculated plates are incubated overnight and the MIC of surviving colonies determined. This method has been applied to two species, S. pneumoniae and P. aeruginosa, for several fluoroquinolones (Table 3) (47,48).

Moxifloxacin exceeds the mutant prevention concentrations for S. pneumoniae, and ciprofloxacin exceeds the mutant prevention concentration for P. aeruginosa (both, 2 mg/L) by achieving maximum serum concentrations of 4.5 mg/L and 3.0 mg/L, respectively. These serum concentrations significantly exceed the mutant prevention concentrations; therefore, these agents are postulated to prevent mutant selection of S. pneumoniae and P. aeruginosa, respectively. Levofloxacin does not achieve mutant prevention concentrations of 8 mg/L in serum and thus may not inhibit mutant selection (47,48).

Approval of ciprofloxacin in the United States in 1987 was accompanied by its rapid inclusion onto most hospital formularies. Initial use was predominantly for P. aeruginosa and other problematic gram-negative infections. However, after ofloxacin was introduced in 1992, some formularies substituted this drug for ciprofloxacin on the basis of cost alone. Similarly, levofloxacin was approved by the U.S. Food and Drug Administration in 1997–1998 for a broad range of infections and was added to formularies in an effort to reduce costs. The clinical consequences of these substitutions was not apparent at the time; however, the epidemiologic data soon emerged that reflected how varying levels of antimicrobial activity could make an impact on pathogen susceptibility and clinical outcomes.

Peterson and colleagues (49) noted decreases in P. aeruginosa susceptibilities to ciprofloxacin and ofloxacin of 21% and 23%, respectively, from 1992 to 1994. This decrease occurred after their medical center switched from ciprofloxacin to ofloxacin as the primary quinolone. In 1994, ciprofloxacin was reintroduced as the primary quinolone, and a 7% recovery in ciprofloxacin activity to P. aeruginosa was reported within 6 months.

Similarly, Rifenburg et al. (50) assessed the effect of fluoroquinolone usage on P. aeruginosa susceptibilities and collated data from 109 hospitals during 1993 to 1996. Greater use of ofloxacin was associated with lower P. aeruginosa susceptibilities. Bhavnani et al. (51) collected data from 145 hospitals and found a significant correlation between use of ofloxacin, but not ciprofloxacin, and decreasing P. aeruginosa susceptibilities. Additionally, the study suggested a deleterious effect of levofloxacin use on P. aeruginosa susceptibilities (51).

Introduction of levofloxacin in 1998 to replace ciprofloxacin in a tertiary-care university medical center resulted in a significant decrease in P. aeruginosa susceptibilities (from 74% to 57% in a 3-year period) and E. coli susceptibility to ciprofloxacin (from 99% to 89% in a 3-year period). Levofloxacin use rose from 91.2 to 272.8 defined daily dose (DDD)/1,000 patient days (199%) in the 3-year period (52). This volume of usage exceeds that of 50 DDD/1,000 patients, a threshold suggested by Austin et al. (53) as a predictive driver in selecting for antimicrobial resistance during a 2-year period. Zambrano et al. (54), at the same institution, recently reported a significant correlation between increased levofloxacin use and declining fluoroquinolone susceptibilities among ICU isolates of K. pneumoniae (96% to 79% [p<0.008]) and P. aeruginosa (82% to 67% [p<0.01]).

Similarly, another group reported (55) that after levofloxacin was added to the formulary, levofloxacin use as a proportion of total fluoroquinolone use increased from <2% to >22% over a 6-month period (from 3rd quarter 1999 to 1st quarter 2000). During the period of 1st quarter 1998 to 2nd quarter 2000, the susceptibility of P. aeruginosa to ciprofloxacin decreased by 11% (82% to 71%). The use of parenteral antipseudomonal agents such as gentamicin, imipenem, ceftazidime, and piperacillin/tazobactam increased concurrently, suggesting that physicians began using non-fluoroquinolone combination therapy when treating serious gram-negative infections. Furthermore, the antimicrobial cost reductions anticipated from switching to a less expensive fluoroquinolone on formulary were not realized. In 3rd quarter 2000, levofloxacin was replaced with ciprofloxacin as the main gram-negative fluoroquinolone, a substitution associated with a subsequent 6% increase in ciprofloxacin activity against P. aeruginosa during the next year.

Because the ICU has been a focal point of antimicrobial resistance, the Centers for Disease Control and Prevention initiated Project ICARE in 1996 (56). Specific data regarding fluoroquinolone use and fluoroquinolone susceptibility among P. aeruginosa isolates were presented for the period 1996–1999 by Hill et al. (57). No correlation was found between prevalence of quinolone resistance and total use of ciprofloxacin/ofloxacin. However, significant associations were found between fluoroquinolone resistance and combined use of ciprofloxacin, ofloxacin, and levofloxacin (p<0.019) and by use of levofloxacin alone (p<0.006) (57).

Likewise, recent studies suggest that using a less potent fluoroquinolone against S. pneumoniae for treating community and hospital respiratory tract infections may be affecting the activity of all fluoroquinolones against this respiratory pathogen (58–60). Fluoroquinolone resistance in S. pneumoniae has been reported, most notably from Hong Kong (58). A 1998 study of 181 S. pneumoniae isolates showed resistance to ciprofloxacin in 12.1%, to levofloxacin in 5.5%, and to trovafloxacin in 2.2%. By early 2000, levofloxacin resistance had increased to include 13.3% of all S. pneumoniae and 27.3% among penicillin-resistant strains (59). These strains also demonstrated elevated MICs to newer class members such as gatifloxacin (12.8%) and moxifloxacin (12.2%) (the latter occurring exclusively in highly penicillin-resistant strains). Additionally, fluoroquinolone resistance appears to be emerging in other countries such as Canada, where resistance rates have increased from 0% in 1993 to 1.7% in 1997/1998 combined (60).

Inappropriate use of antimicrobial agents has been associated with adverse consequences, including therapeutic failure, development of resistance, and increased health-care costs. One example of a mismatch between pharmacodynamics and clinical infection was in the use of ciprofloxacin for community-acquired pneumonia. The pharmacodynamics of the dose typically prescribed in these cases (ciprofloxacin 250 mg b.i.d.) are inappropriate for treating pneumococcal pneumonia, especially in seriously ill patients (41). By 1994, approximately 15 cases of S. pneumoniae infections that did not respond to ciprofloxacin had been reported, primarily in seriously ill patients and associated with contraindicated medications and other important medical issues (61). These events prompted the U.S. Food and Drug Administration to modify the package insert to warn against empiric use of ciprofloxacin for respiratory infections in which S. pneumoniae would be a primary pathogen. Consequently, ciprofloxacin has been used less frequently in these types of infections.

By contrast, >50% of levofloxacin use has been for the treatment of respiratory infections. Since 1999, at least 20 case reports of pulmonary infections that did not respond to levofloxacin therapy have been published (25,62–69). Three of the patients died due to fulminant pneumococcal infections that were unresponsive to levofloxacin therapy at approved dosage (25,62,69). Very few of these cases were in immunosuppressed patients. Reports of pneumococcal failures on the standard dosage of levofloxacin, 500 mg every 24 h, have also been described in two clinical trials, one in a patient with acute exacerbation of chronic bronchitis and the other in a patient with community-acquired pneumonia (70,71) (Table 4). In some of the 21 case reports, the treatment failed, and the pathogen developed levofloxacin resistance during therapy.

Davidson et al. (25) recently published details of four cases of pneumococcal pneumonia in which levofloxacin therapy failed. Two of the patients had no history of prior fluoroquinolone use and were levofloxacin susceptible beginning therapy, but their S. pneumoniae isolates were levofloxacin-resistant after therapy. These resultant mutants exhibited increased MICs to the newer fluoroquinolones moxifloxacin and gatifloxacin as well, thus decreasing those agents’ potential effectiveness. Both Weiss et al. (72) and Ho et al. (73) demonstrated clear risk factors (Table 5) associated with the development of fluoroquinolone resistance, including prior exposure of the patient to first- or second-generation fluoroquinolones (i.e., ciprofloxacin, levofloxacin, and ofloxacin) and history of chronic obstructive pulmonary disease.

The fluoroquinolone class of antimicrobial agents is being increasingly used empirically as resistance has developed to the more traditional antimicrobial agents. Guidelines now recommend fluoroquinolones as first-line empiric therapy for urinary tract infections in regions were trimethoprim/sulfamethoxazole resistance is >10% to 20% (28), and fluoroquinolones are recommended as alternative empiric regimens in some patients with community-acquired pneumonia (26,27). Though increased use of these agents would be expected to lead to increased resistance, a targeted approach to fluoroquinolone prescribing, emphasizing their appropriate use, may reduce development of antimicrobial resistance and maintain class efficacy.

Evidence is mounting that suggests a link between inappropriate fluoroquinolone use, development of antimicrobial resistance against the entire fluoroquinolone class, and clinical failure. To maintain the activity of the fluoroquinolone class, clinicians need to implement an evidence-based approach to antimicrobial selection, particularly a strategy in which the most pharmacodynamically potent fluoroquinolone is matched, on an empiric basis when required, to anticipated bacterial pathogens.

Three major factors are associated with increasing resistance to fluoroquinolones (74): 1) underdosing, i.e., use of a marginally potent agent whose MIC is barely reached in serum or infected tissues; 2) overuse of agents known to encourage resistant mutants; and 3) the inability to readily detect and respond to changes in antimicrobial susceptibilities. Traditional reporting of susceptibility data may be misleading and may not readily identify initial changes in resistance patterns or differences between agents of the same class.

To preserve fluoroquinolone activity, the activity of these agents must be continually assessed, and these agents must be used appropriately. The individual attributes of a given drug should be matched with the likely pathogen at specific infective sites. Expecting a single fluoroquinolone to be suitable for all infections is unreasonable, and excessive use of any single fluoroquinolone for all indications will lead to resistance that will adversely affect the entire class.

Given the defined strategy of selecting the agent with the best pharmacokinetic and pharmacodynamic profile against the known or suspected pathogen, an appropriate therapeutic choice for most serious infections, such as nosocomial pneumonia in which P. aeruginosa is a known or suspected pathogen, would currently include ciprofloxacin in combination with an antipseudomonal β-lactam or an aminoglycoside antibiotic. This recommendation is based on the lower MIC₉₀ and mutant prevention concentrations for this fluoroquinolone against P. aeruginosa and higher Cmax/MIC and AUC/MIC ratios compared to other members of the class. Likewise, for most other gram-negative infections of the skin and urinary tract, including P. aeruginosa infections, ciprofloxacin monotherapy is appropriate.

Ciprofloxacin, levofloxacin, and gatifloxain all achieve high concentrations in urine; thus, they would all be appropriate choices for treating urinary tract infections in the community. Ciprofloxacin would be the most appropriate therapy in cases where P. aeruginosa is a known or suspected pathogen. For other gram-negative infections, levofloxacin or gatifloxacin should be prescribed in appropriate doses to surpass the mutant prevention concentrations at the infection site.

For infections in which S. pneumoniae is anticipated to be the most likely pathogen (e.g., community-acquired pneumonia), moxifloxacin, which currently has the best antipneumococcal pharmacodynamic activity and the lowest mutant prevention concentrations against this organism, would represent a prudent therapeutic choice. By contrast, levofloxacin MIC₉₀s against S. pneumoniae are significantly higher than those of moxifloxacin and gatifloxacin. The AUC/MICs and Cmax/MICs are also lower for levofloxacin against S. pneumoniae, and serum concentrations of a standard dose of levofloxacin for community-acquired pneumonia do not reach the mutant prevention concentrations for S. pneumoniae. For these reasons, levofloxacin may not be the best choice for infections caused by S. pneumoniae. Furthermore, recent reports of levofloxacin failures in cases of community-acquired pneumonia caused by S. pneumoniae evoke concern.

The targeted strategy proposed in this review is being implemented in a variety of institutions since the introduction of the third-generation fluoroquinolones. Documenting the effect of this approach on hospital susceptibilities over time will be important. Additionally, susceptiblities in these hospitals need to be compared to those in hospitals that use a single fluoroquinolone more broadly.