Efficacy of High Doses of Levofloxacin in Experimental Foreign-Body Infection by Methicillin-Susceptible Staphylococcus aureus

Microorganism used and determination of MICs and minimum bactericidal concentrations (MBCs).We used methicillin-susceptible S. aureus strain ATCC 29213, which is also susceptible to all antibiotics tested in our experiments.

The MIC and MBC were determined according to standard recommendations (31). We used a Mueller-Hinton broth (MHB) macrodilution method with a final inoculum of 10⁵ to 10⁶ CFU/ml under exponential growth conditions. Both MICs and MBCs were determined after 24 h of incubation at 37°C; the MIC was defined as the minimum concentration of antibiotic that was able to inhibit visible bacterial growth, while the MBC was the lowest concentration which killed 99.9% of the original inoculum.

We also determined MBCs during the stationary phase of growth. Bacteria, which were recovered from an overnight culture in Trypticase soy broth, were centrifuged and resuspended in a nutrient-restricted medium (phosphate-buffered saline [PBS], 1% glucose, and 4% MHB), thus ensuring that bacteria remained stable for up to 24 h under these conditions (48). A macrodilution method with a high inoculum of 10⁸ CFU/ml was used. The MBCs were defined as described above.

In vitro time-kill curves.Time-kill curves to determine bactericidal activity during the exponential growth phase were derived following standard recommendations (30), with 10 ml of MHB, a final inoculum of 10⁵ CFU/ml, and a prefixed concentration of antibiotic (multiples of the MIC, designated 0.5×, 1×, 2×, 4×, 8×, 16×, 128×, 256×, and 512× MIC, according to the drug used). Quantitative bacterial counts were determined as log CFU/ml at 8 and 24 h of incubation at 37°C. To avoid carryover antimicrobial agent interference, the sample was placed on the plate in a single streak down the center and allowed to absorb into the agar until the plate surface appeared dry, and the inoculum was then spread over the plate.

We also carried out studies with bacteria in the stationary phase by using a final inoculum of 10⁸ CFU/ml and replacing MHB with a nutrient-restricted medium as described above. The prefixed antibiotic concentrations were equivalent to peak, trough, and intermediate levels in tissue cage fluids (TCF); the concentrations (μg/ml) tested were 32, 16, and 2 for cloxacillin; 8, 4, and 1 for levofloxacin; 16, 4, and 2 for linezolid; 8 and 4 for rifampin, and 32, 16, and 4 for vancomycin.

The reduction in log CFU/ml counts at the end of experiments with respect to the initial inoculum (log CFU/ml) was determined for each antibiotic concentration, and efficacy was expressed as the percentage of eradicated bacteria (PEB) in both the exponential and stationary studies.

Preparation of inoculum for in vivo use.Bacteria from overnight cultures on 5% blood agar plates were grown for 4 to 6 h in Trypticase soy broth. They were then centrifuged and resuspended in sterile saline solution adjusted to an optical density of a 0.5 McFarland standard and finally diluted to a concentration of 0.2 × 10⁶ to 2 × 10⁶ CFU/ml.

Antimicrobial agents.For all in vitro experiments, and for the in vivo experiment with levofloxacin, the purified powder of each antibiotic was resuspended following the respective laboratory's recommendations. For the remaining in vivo studies, we used commercial products, with the necessary dilutions being performed to achieve a final volume that could be administered to animals.

All antibiotics were supplied by their respective laboratories, as follows: levofloxacin was supplied by Aventis Pharma, Frankfurt, Germany; rifampin by Aventis Pharma, Madrid, Spain; linezolid by Pfizer, Madrid, Spain; and cloxacillin and vancomycin by Normon, Madrid, Spain.

Animal studies. (i) Animal model.The study was previously approved by the Ethical Committee for Animal Experiments at the University of Barcelona.

Male Wistar rats weighing 220 to 250 g at the beginning of the experiments were used; they were given food and water ad libitum throughout the study. For the surgical procedure, the rats were anesthetized with an intraperitoneal injection of ketamine (80 mg/kg; Parke-Davis, Madrid, Spain) plus xilacyne (20 mg/kg; Bayer, Barcelona, Spain); they also received analgesia with buprenorphine (0.01 mg/kg; Schering Plough, Madrid, Spain). To obtain samples during the experiment, rats were anesthetized with isofluorane (Abbott, Madrid, Spain). Animals were sacrificed by cardiac puncture with thiopental (Braun, Barcelona, Spain).

We used a model previously described by Lucet et al. (26), the only modification being the number of tissue cages for each animal. In brief, two multiperforated polytetrafluoroethylene (Teflon) tissue cages (internal and external diameter of 10 and 12 mm, respectively; length, 32 mm) containing two polymethylmethacrylate coverslips (7 by 7 by 1 mm) each were subcutaneously implanted in rats. Three weeks after surgery, TCF were percutaneously obtained and checked for sterility; when bacterial contamination was excluded, TCF were then inoculated with 0.1 ml of saline solution containing an inoculum of S. aureus. Three weeks after inoculation (designated day 1), 0.1 ml of TCF was processed (see below) to determine bacterial counts, which were expressed as log CFU/ml; those with fewer than 10⁵ CFU/ml were excluded. Animals were sacrificed only if results for both tissue cages were not valid. Rats included in the experiment were then assigned randomly to groups to be intraperitoneally treated for seven days with antibiotics (twice daily for all drugs except levofloxacin, which was administered once daily) or to be left untreated. Twelve to 24 h after the end of therapy (designated day 8), TCF were cultured to determine the bacterial counts. The difference in the number of log CFU/ml between day 1 and day 8 (Δlog CFU/ml) was defined as the criterion of antibiotic efficacy. The in vivo PEB from these counts with respect to initial counts was also determined as described above for in vitro studies.

Prior studies to evaluate the spontaneous course of tissue cage fluid infection were performed by analyzing TCF on days 7, 14, 16, 21, 28, and 35 after inoculation. After TCF was infected with 0.2 × 10⁶ to 2 × 10⁶ CFU/ml of initial inoculum, lower bacterial counts (means of log CFU/ml ± standard deviations) during weeks 1 and 2, ranging from 4.86 ± 0.46 on day 7 to 5.67 ± 0.35 on day 14, were observed. Bacterial counts close to that of the initial inoculum were recovered on day 16 (6.39 ± 0.8) and after a chronic and stable infection was established, as previously reported (26), ranging from 7.07 ± 0.5 on day 20 to 7.2 ± 0.9 on day 35. Spontaneous shedding of approximately 15 to 20% of the tissue cages was noted.

(ii) Processing of TCF and coverslips.The processing schedules were performed according to previous reports, in which they were described as being harmless to bacteria (9, 26). TCF were obtained by percutaneous puncture under aseptic conditions and were sonicated (150 W for 1 min; Afora, Madrid, Spain) to disrupt bacterial clumps. A sample of 100 μl of the sonicated fluids and their 10-fold dilutions was plated on a Trypticase soy agar plate with 5% sheep blood for 48 h at 37°C; to avoid a carryover antimicrobial effect, we proceeded as described above for kill curves. Bacterial counts were recorded as log CFU per ml. The lower detection limit was 100 CFU/ml.

When animals were sacrificed, coverslips from tissue cages were removed under aseptic conditions and rinsed three times in 1 ml of PBS; they were then incubated in 1 ml of PBS with trypsin (6 U/ml; Sigma, Madrid, Spain) for 20 min at 37°C. Finally, the remaining PBS was sonicated to recover adherent bacteria, with the final fluid being used to screen resistant bacteria (see below).

Pharmacokinetic studies.Prior to therapeutic experiments, pharmacokinetic (PK) and PD profiles of each drug in healthy animals were studied. We administered a single weight-adjusted dose of antibiotic to a group of 10 rats, and samples of blood and TCF were obtained at different time points (after 30 min and 1, 2, 3, 4, 6, 8, 12, and 24 h, according to the antibiotic). In all cases, a minimum of four samples for each time point at a minimum of six time points for each drug were studied. Moreover, peak and trough concentrations were determined on day 4 of therapy in TCF to test the equilibrium concentration achieved. Blood and TCF samples were obtained by cardiac and percutaneous puncture, respectively, through the tissue cage; both samples were then centrifuged, and the remaining serum and fluid samples were conserved at −70°C until analysis.

The PK-PD parameters studied for both serum and TCF were as follows: the peak concentration (C_max), the elimination half-life (t_1/2), the elimination rate constant, the area under the concentration-time curve (AUC [μg · h/ml]) over 24 h, the time the drug concentration remained above the MIC (T_>MIC), the AUC/MIC ratio, and the C_max/MIC ratio.

Based on previous experimental studies (13, 33, 43, 40), the definitive antibiotic doses were those achieving PD parameters in TCF that were similar to human ones in serum with conventional doses. For all drugs except cloxacillin, we adjusted AUC values to obtain similar AUC/MIC ratios in animals and in humans (7, 8, 24, 25, 32, 34). Free-drug concentrations were used for all drugs except vancomycin, for which the total drug and its protein binding were measured.

Final doses of antibiotics used were as follows: cloxacillin, 200 mg/kg/12 h; linezolid, 35 mg/kg/12 h; rifampin, 25 mg/kg/12 h; vancomycin, 50 mg/kg/12 h; and levofloxacin, 50 mg/kg/24 h (levofloxacin 50) and 100 mg/kg/24 h (levofloxacin 100).

Antibiotic assays.A bioassay method (5) was used to determine drug concentrations of all antibiotics except vancomycin. The following microorganisms were used in assays: S. aureus ATCC 29213 for cloxacillin, Staphylococcus epidermidis ATCC 27626 for rifampin, Bacillus subtilis ATCC 12432 for linezolid, and Escherichia coli ATCC 35218 for levofloxacin. Studies were performed according to the previous literature.

The vancomycin concentrations were determined by fluorescent polarization immunoassays using a TDX analyzer (Abbott, Madrid, Spain). To measure the binding of vancomycin in TCF, ultrafiltration was performed using a centrifugal filter device (Centrifree; Millipore Corp., Bedford, Mass.) (42). In vivo TCF samples containing vancomycin were analyzed by fluorescent polarization immunoassays; samples were first ultrafiltrated or were used without filtration to determine the protein binding. PBS-buffered samples containing vancomycin was used as controls.

Resistance to antimicrobial agents.For the linezolid and rifampin therapeutic groups, the development of resistance at the end of therapy was screened. In all cases, 100 μl of direct TCF and processed fluid from coverslips were cultured in plates containing 4 mg/liter linezolid or 4 mg/liter rifampin. Results were expressed qualitatively as positive (with any macroscopic growth) or negative (with no macroscopic growth).

Statistical studies.All bacterial counts are presented as log CFU/ml (means ± standard deviations). Data were found to be normally distributed when the Kolmogorov-Smirnov test was applied. Analysis of variance and Scheffe's correction were used to compare differences between groups in bacterial counts. Studies of correlation were performed using Pearson's coefficient from mean PEBs. For all tests, differences were considered statistically significant when P values were <0.05.

In vitro studies.The MICs and MBCs for the exponential and stationary phases for each antibiotic are shown in Fig. 1.

• Open in new tab
• Download powerpoint

FIG. 1.

Relationships between in vivo, in vitro log-phase, and in vitro stationary-phase PEBs. Drug concentrations (μg/ml) for the log phase and stationary phase, respectively, are as follows: cloxacillin (OXA), 4 and 32; levofloxacin (LVX), 4 and 4; linezolid (LZD), 16 and 16; vancomycin (VAN), 16 and 32; and rifampin (RIF), 8 and 8. MICs (μg/ml) for each antibiotic in the log phase are noted. MBCs (μg/ml) for the log phase and stationary phase are shown on top of the respective column for each antibiotic. Errors bars indicate standard deviations.

In log-phase studies, cloxacillin, vancomycin, and levofloxacin reached a MBC/MIC ratio of 2 as the bactericidal antibiotic profile, whereas linezolid and rifampin had a ratio of 16 and >512, respectively, in terms of bacteriostatic activity. In comparison with log-phase studies of MBCs, stationary-phase studies revealed increases in the MBCs of all antibiotics; MBCs were found at clinically relevant concentrations only for levofloxacin. Rifampin showed high MBCs in both phases.

Time-kill curves for the exponential phase showed bactericidal activity by levofloxacin, cloxacillin, and vancomycin at concentrations of 2× MIC; observations at higher concentrations showed that levofloxacin and cloxacillin were able to kill all original inoculum at 2× MIC and 8× MIC, respectively. At 8 h, levofloxacin reached a potent killing rate, one that was even better than that of cloxacillin for all equivalent concentrations tested.

Stationary-phase kill curve studies showed that only levofloxacin reached bactericidal activity, although higher concentrations (8× MIC) than those observed during the exponential phase were required. Rifampin never achieved a bactericidal effect in either phase, even when concentrations as high as 512× MIC were tested; some log-phase studies resulted in final regrowth due to the development of resistance, whereas this phenomenon was not observed for the stationary phase. Linezolid showed bactericidal activity at high doses (16× MIC) for exponential-phase studies but only bacteriostatic activity in stationary-phase kill curves. Time-kill curves for the clinically relevant concentrations for which the highest efficacy was achieved during the two phases are shown in Fig. 2; we noted that the concentrations required in the stationary-phase studies were close to the peak level for all antibiotics, whereas lower concentrations for some antibiotics were required in the log phase.

• Open in new tab
• Download powerpoint

FIG. 2.

Time-kill curves for the log phase (a) and stationary phase (b) with clinically representative concentrations (μg/ml). Errors bars indicate standard deviations. Abbreviations: OXA, cloxacillin; VAN, vancomycin; LZD, linezolid; LVX, levofloxacin; RIF, rifampin.

Pharmacokinetic and pharmacodynamic studies.The main parameters in serum and TCF for each antibiotic are shown in Table 1. Two doses of levofloxacin were selected in order to mimic human pharmacodynamic profiles of 500 mg/day (50 mg/kg/day in rats) and 750 to 1,000 mg/day (100 mg/kg/day in rats). The C_max in TCF for cloxacillin was 43 μg/ml. The AUC, C_max/MIC ratio, and AUC/MIC ratio were not determined due to the time-dependent killing of beta-lactams; thus, pharmacodynamic parameters were optimized, allowing a T_>MIC of 100% in TCF. For the remaining antibiotics, the AUC values were consistent with those obtained for humans with conventional doses, and the serum and TCF results were similar for each antibiotic. All antibiotics except linezolid achieved a T_>MIC of 100%. The trough concentrations were calculated on day 4 of therapy for TCF only. Our results didn't show any remarkable differences between the trough concentrations determined on day 4 of therapy and those determined by pharmacokinetic studies with healthy animals; these concentrations (in μg/ml) were, respectively, 3.88 and 3.59 for cloxacillin (the accumulation ratio [trough concentration on day 4/trough concentration on day 1] was 1.08), 1.1 and 0.74 for levofloxacin 100 (accumulation ratio, 1.48), 0.6 and 0.4 for levofloxacin 50 (accumulation ratio, 1.5), 2.6 and 2.4 for linezolid (accumulation ratio, 1.08), 3.8 and 3.4 for rifampin (accumulation ratio, 1.11), and 4.6 and 4.2 for vancomycin (accumulation ratio, 1.09).

Animal studies.A total of 80 rats were used in the experiments and, after tissue cages with inadequate bacterial counts were excluded, the remaining animals were randomized to be treated. The final number of tissue cages for each therapeutic group was as follows: cloxacillin, 19; levofloxacin 50, 20; levofloxacin 100, 19; linezolid, 18; rifampin, 17; vancomycin, 19; and controls, 18.

Quantitative bacterial counts for different groups on day 1 (means ± standard deviations) were 6.44 ± 0.80 for the cloxacillin group, 6.42 ± 0.89 for levofloxacin 100, 6.40 ± 0.80 for levofloxacin 50, 6.82 ± 0.86 for linezolid, 6.91 ± 0.78 for rifampin, 7.03 ± 0.86 for vancomycin, and 6.86 ± 0.88 for the controls; there were no statistical differences between groups on day 1 (P > 0.05).

A comparison of the final decreases in log CFU/ml between groups is shown in Fig. 3. All therapeutic groups were better than the controls (P < 0.05); levofloxacin 100 was the most effective therapy, but it was significantly better than only linezolid (P = 0.03), which showed the least efficacy. Rifampin was the second most active therapy, although resistant strains were detected in 10 of the 17 samples. The screening of development of resistance was performed with the fluid from coverslips and from TCF; resistant strains were detected in fluid from all 10 coverslip samples and in 9 of 10 TCF samples. No resistance developed in the linezolid group at the end of therapy.

• Open in new tab
• Download powerpoint

FIG. 3.

Comparison of decreases in log CFU/ml (means) between groups at the end of therapy. Errors bars indicate standard deviations. Abbreviations: OXA, cloxacillin; VAN, vancomycin; LZD, linezolid; LVX, levofloxacin; RIF, rifampin. *, P < 0.05 versus the control; **, P = 0.03 versus LZD and P < 0.05 versus the control.

The correlations between in vivo and in vitro PEBs for the exponential and stationary studies (at equivalent clinical peak concentrations) were r values of 0.32 and 0.63, respectively. The major discrepancy was due to the rifampin group, in which there was a lack of correlation between in vivo and in vitro efficacy for the two phases. Finally, when the rifampin group was excluded, the correlations (r values) were 0.7 for the exponential phase and 0.96 (P < 0.05) for the stationary phase.