INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

With increasing antimicrobial resistance reported from all over the world, it is important to find antibiotics with enhanced activity against resistant organisms (1). Antibiotic resistance is a particular problem in Enterococcus spp., with some strains showing high-level resistance to aminoglycosides, beta-lactams, and vancomycin. Vancomycin-resistant Enterococcus faecium is being isolated with increasing frequency around the world, particularly from hospitalized patients (5, 12). This organism is quite difficult to eradicate from infected heart valves and from other infected sites (12, 19). Although some success in early trials has been obtained with the new antibiotic combination quinupristin-dalfopristin (Synercid), the search for antibiotics with activity against these organisms continues.

Trovafloxacin is a new fluoronaphthyridone with a long elimination half-life and enhanced activity against gram-positive cocci, including penicillin-susceptible and -resistant strains of Streptococcus pneumoniae, Staphylococcus aureus, Streptococcus pyogenes, etc. Its activity against Enterococcus faecalis and E. faecium is more variable but still greater than those of older fluoroquinolones (6).

The in vitro bactericidal activity of trovafloxacin against Enterococcus spp. was studied with time-kill assays and in an in vitro pharmacokinetic-pharmacodynamic model of infection that allows for the exposure of bacteria to changing concentrations of antibiotic that mimic human pharmacokinetics. Some strains also were studied with ampicillin-sulbactam alone and in combination with trovafloxacin.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Twenty-one strains of vancomycin-resistant E. faecium and two strains of E. faecalis (one of which was vancomycin resistant) were studied (kindly supplied by John Boyce, Antone Medeiros, Leonard Mermel, Lance Peterson, and Steven Opal). Time-kill studies were performed in 10 ml of brain heart infusion broth, and each strain was studied with trovafloxacin alone, with ampicillin-sulbactam alone, and with the combination. An initial inoculum of 10⁶ CFU/ml was used, and antibiotic concentrations approximated levels in serum following a single 300-mg dose of trovafloxacin (3 mg/liter) and a single 2/1-g dose of ampicillin-sulbactam (100/50 mg/liter). These mixtures were incubated at 37°C and were sampled at 0, 4, 8, and 24 h for bacterial counts, which were performed with serial 1:10 dilutions of 20-µl volumes plated in triplicate on antibiotic-free plates (Mueller-Hinton agar supplemented with 5% sheep blood). Bacterial counts are expressed as change of log CFU at 24 h.

Six strains of E. faecium (five of six were vancomycin resistant) also were studied in an in vitro two-compartment dynamic model that mimics human pharmacokinetics (2, 3). This model includes a central compartment that mimics concentrations in serum and a peripheral compartment consisting of hollow-fiber artificial-capillary units placed in series (Bioreactor chambers). Although the probable oral dose of trovafloxacin is likely to be 200 mg per day, in this model trovafloxacin was simulated as 300-mg doses q12h and ampicillin-sulbactam was simulated at 2/1 g q6h intravenously. The same doses were simulated when the drugs were studied in combination. In these studies, log-phase bacteria were used at initial inocula that ranged from 5.7 × 10⁵ to 1.7 × 10⁶ CFU/ml. The peripheral compartments were sampled for cell counts q2h for 30 h (except hour 16). The central compartment was sampled seven to eight times following the first dose and five to seven times following subsequent doses for bioassay determinations of drug concentrations.

Trovafloxacin concentrations in the central and peripheral compartments were determined by bioassay. Bacillus subtilis (ATCC 6633)-seeded plates containing Antibiotic Medium no. 1 (Difco) were used. Wells were made in the agar and filled with 25 ml of sample. Before the wells were filled, Penase (Difco Penase concentrate [penicillinase, 10,000,000 IU/ml]) was added to all samples that contained both trovafloxacin and ampicillin-sulbactam in order to inactivate the ampicillin. Plates were incubated overnight at 37°C, and zone sizes were measured with calipers the following day.

Trovafloxacin concentrations were calculated against a four-point standard curve by linear regression. The calibration plots were linear within the range of 0.25 to 1.2 mg/ml. The r² for standard curves ranged from 0.98 to 0.99 (median, 0.99). The median percent error in the standard curves was 1.9% (range, 9.8 to 12.2). The median coefficient of variation for standards and unknown drug concentrations was 8.52% (range, 0 to 24.6) when three to four replicates per sample were studied.

Ampicillin concentrations in the central and peripheral compartments were determined by a well-plate bioassay (14) using the same B. subtilis-seeded plates as for trovafloxacin determinations. The median coefficient of variation for standards and unknown drug concentrations was 16.2, with a range of 0 to 28.7 when three or four replicates per sample were studied. Sulbactam did not interfere with ampicillin determinations at concentrations between 4.7 and 75 mg/ml.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

MICs and kill curve results. The MICs of trovafloxacin and ampicillin-sulbactam for the 21 strains of E. faecium are shown in Table 1. Most of the organisms were resistant to ampicillin and to ampicillin-sulbactam. Trovafloxacin MICs ranged from 0.5 to 32 mg/liter. When E. faecium strains were divided by their susceptibilities to trovafloxacin, those strains for which MICs were 2 mg/liter were more likely to exhibit a 2-log-unit or greater reduction in CFU per milliliter after exposure to 3 mg of trovafloxacin/liter than were strains for which MICs were higher (Table 2).

View larger version (50K): [in this window] [in a new window]   FIG. 1.   Time-kill curves of trovafloxacin (TF), ampicillin-sulbactam (AMP/SUL), and the combination (AMP/SUL + TF) for a trovafloxacin- and ampicillin-sulbactam-resistant strain (A) and for a trovafloxacin-susceptible strain (B) of vancomycin-resistant E. faecium.

Results in the in vitro model. The concentrations of each antibiotic in the central compartment as measured by bioassay are shown in Fig. 2. Reasonable approximations of human concentration-time curves were achieved in the model (17). Data for four strains of vancomycin-resistant E. faecium tested in the in vitro dynamic model are shown in Fig. 3. Two of these strains, 359 and 5924, were susceptible to trovafloxacin (MICs, 2 mg/liter), and the MICs for the other two, 391 and 4162, were higher (4 mg/liter). Both of the sensitive strains showed a 2- to 4-log-unit reduction in bacterial counts, with a maximal effect at 10 to 18 h (Fig. 3). Ampicillin-sulbactam alone produced inhibition with one strain, but the other three strains were neither inhibited nor killed.

View larger version (53K): [in this window] [in a new window]   FIG. 2.   Concentrations of trovafloxacin (A) and ampicillin (B) in the central compartment in in vitro model experiments.

View larger version (38K): [in this window] [in a new window]   FIG. 3.   Effects of trovafloxacin (TF) (300 mg q12h), ampicillin-sulbactam (AMP/SUL) (2/1 g q6h), and the combination against four strains of E. faecium studied in an in vitro pharmacokinetic-pharmacodynamic model.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

This paper shows that trovafloxacin simulating a dose of 300 mg q12h (which is higher than the proposed daily oral dose of 200 mg) has modest bactericidal activity against strains of vancomycin-resistant E. faecium if the MICs for these strains are 2 mg/liter. While it is unlikely that trovafloxacin or any of the new fluoroquinolone antibiotics will have major roles in the treatment of infections caused by vancomycin-resistant E. faecium, 9 of 21 (43%) strains studied here were inhibited by 2 mg of trovafloxacin/liter. All nine of these strains showed 1- to 3-log-unit reductions in viable counts at 24 h in time-kill assays. This study also shows the utility of in vitro dynamic models that allow for multiple dosing. As seen in Fig. 3, simulated trovafloxacin dosing (albeit at the higher, 300-mg dose) did result in 2- to 4-log-unit killing of susceptible strains, although some regrowth was apparent beginning at 22 h.

Other investigators have reported that 30 to 81% of vancomycin-resistant E. faecium strains are susceptible in vitro to trovafloxacin at concentrations of 2 mg/liter (9, 10). Several recent studies also suggest that trovafloxacin might have some useful activity against vancomycin-resistant E. faecium in a variety of models. Using agar dilution, Coque and colleagues reported trovafloxacin MICs for 16 strains of vancomycin-resistant E. faecium ranging from 0.12 to 8 mg/liter, and 81% of these strains were susceptible to 2 mg/liter (10). Using 3 mg of trovafloxacin/liter in time-kill assays, they reported reductions of 2 or more log units at 24 h with five of six strains of vancomycin-resistant E. faecium (10). Also, Cormican and Jones (9) used broth microdilution and disk diffusion methods to study trovafloxacin activity against 131 vancomycin-resistant enterococci and reported potentially useful activity against 30 to 37% of tested strains.

The MIC at which 90% of isolates are inhibited (MIC₉₀) of trovafloxacin was reported to be 0.5 mg/liter for 29 enterococci unidentified at the species level in a recent report (7), and Bonilla et al. reported a trovafloxacin MIC₅₀ of 2 mg/liter (4). Another group reported trovafloxacin MIC₉₀s of 2 mg/liter against 10 strains of E. faecium, 16 mg/liter against 15 ciprofloxacin-resistant strains of E. faecium, and 0.25 mg/liter against 20 strains of E. faecalis (8). Endtz and colleagues reported that trovafloxacin was active against 20 strains of vancomycin-resistant E. faecalis, with MICs ranging from 0.125 to 2 mg/liter (13). Also, trovafloxacin was found to be active in media against three vancomycin-susceptible and three vancomycin-resistant E. faecium strains, but less activity was found intracellularly against vancomycin-resistant E. faecium (15).

Consistent demonstration of synergism or antagonism with ampicillin-sulbactam was not obtained in these studies. However, careful examination of Fig. 3 suggests that for two of the four tested strains, the addition of ampicillin-sulbactam might inhibit the late bacterial regrowth seen with trovafloxacin, despite rather unimpressive activity of the beta-lactam alone. Coque et al. (11) also showed that combinations of trovafloxacin with ampicillin-sulbactam did not show enhanced activity except against one strain of E. faecium. Another report suggested that trovafloxacin and ampicillin-sulbactam were synergistic or partially synergistic against 10 strains of enterococci unidentified at the species level when tested with an agar dilution checkerboard method (16).

Although some anti-vancomycin-resistant E. faecium activity has been demonstrated in vitro and in some experimental models of infection, it is unlikely that trovafloxacin will be uniformly effective in eradicating these organisms from patients with serious infections. In addition, the reported 70% binding to serum protein (18) might further reduce its activity. Nonetheless, for some select patients infected with E. faecium strains for which MICs are low (2 mg/liter), trovafloxacin in high doses with or without other antibiotics might have a beneficial effect. Also, the possible suppression of late bacterial regrowth by the combination of trovafloxacin plus ampicillin-sulbactam seen with some strains in the in vitro kinetic model presented here is of interest. The demonstrated potential activity of trovafloxacin against some vancomycin-resistant E. faecium strains suggests that further study with clinically infected patients may be of benefit.