Bactericidal Activities of Meropenem and Ertapenem against Extended-Spectrum-β-Lactamase-Producing Escherichia coli and Klebsiella pneumoniae in a Neutropenic Mouse Thigh Model^▿

Antimicrobial test agents.

Standard analytical-grade meropenem (Astra Zeneca, Wilmington, DE) and ertapenem (Merck, West Point, PA) were used for in vitro testing. For in vivo analysis, commercially available meropenem and ertapenem for injection were obtained from their respective drug manufacturers. One gram of meropenem powder (Astra Zeneca) and 1 g of ertapenem powder (Merck) were reconstituted with 20 ml and 10 ml, respectively, of sterile water for injection as per the manufacturer's instructions. Final concentrations were diluted to achieve the desired milligram-per-kilogram of body weight doses administered immediately prior to each experiment. These final solutions were kept under refrigeration when not in use and were discarded 24 h after reconstitution.

Bacterial isolates.

Thirty-one clinical isolates of ESBL-producing E. coli or K. pneumoniae were included. These isolates were either collected from the Hartford Hospital microbiology department or collected as part of the MYSTIC database in both North and South America from 2004 and 2005 (29). Additionally, two non-ESBL E. coli isolates were tested for comparison.

Susceptibilities.

The MICs of meropenem and ertapenem were determined using either Etest methodology (AB Biodisk, Solna, Sweden) or the microdilution method according to CLSI guidelines with cation-adjusted Mueller-Hinton broth (20 to 25 mg/liter calcium, 10 to 12.5 mg/liter magnesium) (2a). The MICs of three randomly selected isolates were also tested at a greater starting inoculum of 10⁷ CFU/ml by use of broth microdilution methods. The specific ESBL enzyme was not known for the majority of these clinical isolates.

Thigh infection model.

Specific-pathogen-free, female ICR mice weighing approximately 25 g each were obtained from Harlan Sprague Dawley, Inc. (Indianapolis, IN) and utilized throughout these experiments. The animals were maintained and utilized in accordance to National Research Council recommendations and provided food and water ad libitum. Mice were rendered transiently neutropenic by intraperitoneal injections of cyclophosphamide at doses of 150 and 100 mg/kg of body weight at 4 days and 1 day before inoculation, respectively (11). Three days prior to inoculation, the mice received a single intraperitoneal injection of 5 mg/kg of uranyl nitrate to induce a predictable degree of renal impairment (2).

A suspension of each isolate was prepared from a fresh subculture of organism that had been incubated for less than 20 h and diluted to achieve a final inoculum of 10⁵ or 10⁷ CFU/ml for the standard- and high-inoculum bacterial density studies, respectively. Thigh infection was produced by a single intramuscular injection of 0.1 ml of inoculum into each mouse thigh 2 h prior to the initiation of antimicrobial therapy.

PK studies and dosing regimen determination.

PK studies were performed to find a meropenem regimen in mice that simulated the PK profile observed for healthy adults receiving meropenem at 1 g every 8 h. These regimens were designed to mimic the free drug percent time above the MIC (fT>MIC) observed for humans over a wide range of MICs. The free drug concentration-time profile in humans was derived from the work of Dreetz et al., and the protein binding of 8% in this study was considered to be equivalent for both humans and mice (7, 21). Previously defined meropenem PK parameters for mice (26) were entered into WinNonLin (version 3.3; Pharsight, Mountain View, CA) to determine the approximate dosing frequency to best approximate the human PK profile. Based on these PK parameters, a resultant dosing regimen of 65, 30, 20, and 5 mg/kg of body weight at 0, 1.5, 3, and 5 h in three 8-h intervals over 24 h best approximated the human fT>MIC over a wide MIC range. Prior to the efficacy studies, confirmation of these exposures was performed with a separate group of infected mice.

Meropenem serum samples were analyzed by a validated high-performance liquid chromatography (HPLC) method as previously described (9). The meropenem HPLC assay was linear (r = 0.998) over a concentration range of 0.25 to 40.0 μg/ml. The intraday quality control samples (n = 10) of 0.5 and 30.0 μg/ml had percents coefficient of variation of 4.39% and 3.77%, respectively. The interday quality control samples (n = 6) of 0.5 and 30.0 μg/ml had percents coefficient of variation of 3.75% and 7.44%, respectively.

With regards to ertapenem, previous work from our laboratory determined that 50 mg/kg of body weight every 6 h for mice achieved a concentration-time profile and fT>MIC similar to those observed following ertapenem at 1 g every 24 h for humans (33). Similar to meropenem, a more complex regimen, which closely mimicked the concentration-time profile following the administration of ertapenem at 1 g every 24 h for humans, was investigated. However, since no difference in efficacy was observed between the simplified and complex regimens, the regimen of 50 mg/kg of body weight every 6 h was utilized in this analysis (6). The ertapenem free drug concentration-time profile was derived from the work of Musson et al. (24) and was summarized by Nix et al. (25). An ertapenem protein binding of 95.5% in mice, determined by Xuan et al. (33), was used to determine fT>MICs at various MICs.

Treatment regimens.

Human PK-simulated regimens of meropenem and ertapenem in mice were administered as 0.2-ml subcutaneous injections and studied for each pathogen in groups of three mice over a 24-h period. Control animals received sterile normal saline subcutaneously at the same volume (0.2 ml) and schedule used in the active drug regimen. Untreated control mice (three per group) were sacrificed just prior to antibiotic initiation (0 h) and after 24 h. After the 24-h treatment period, all animals were euthanized by CO₂ exposure followed by cervical dislocation.

Efficacy as assessed by bacterial density.

After sacrifice, the thighs were removed and individually homogenized in normal saline. Serial dilutions were plated on Trypticase soy agar with 5% sheep blood for CFU determination. Efficacy (change in bacterial density) was calculated as the log₁₀ change in bacterial CFU/thigh obtained for the meropenem- or ertapenem-treated mice after 24 h from the preantibiotic CFU/thigh measured for the 0-h control animals.

In vitro susceptibility testing.

Meropenem and ertapenem MICs for the study isolates are displayed in Table 1. Against the 31 ESBL isolates, meropenem displayed greater in vitro potency, with a median MIC of 0.125 μg/ml, than did ertapenem, with a median MIC of 0.5 μg/ml. The MICs for the three isolates that were tested at a higher inoculum of 10⁷, E. coli isolates 315, 321, and 322, were ≥64-fold higher than those at the standard inoculum. Meropenem MICs rose from 0.032, 0.032, and 0.094 μg/ml to 4, 4, and 8 μg/ml, respectively, for these three isolates. Likewise, ertapenem MICs rose from 0.012, 0.012, and 0.032 μg/ml to 1, 1, and 2 μg/ml, respectively.

PK determination.

A comparison of the resultant concentration-time profiles achieved with mice for meropenem and ertapenem to those observed for humans is shown in Fig. 1. The meropenem free area under the curve (fAUC) obtained for mice was higher than that for humans; these values were 61.1 μg·hr/ml and 47.5 μg·hr/ml, respectively. However, as displayed in Table 2, the human and animal fT>MIC values agreed to within 8% over an MIC range of 0.5 to 8 μg/ml. Emphasis was placed on obtaining similar fT>MICs at MICs near the susceptibility breakpoint (i.e., 0.5 to 2 μg/ml for ertapenem and 1 to 4 μg/ml for meropenem). For ertapenem, the fAUC obtained for mice was 27.7 μg·hr/ml, which closely approximated the fAUC of 33.2 μg·hr/ml obtained for humans. The fT>MICs for the mouse and human dosing regimens agreed to within 10% at MICs of ≥0.5 μg/ml (Table 2).

Efficacy as assessed by bacterial density.

The numbers of organisms recovered from the thighs of infected animals serving as 0-h controls (immediately prior to antibiotic treatment) ranged from 4.52 to 5.98 log₁₀ CFU/thigh with a mean ± standard deviation of 5.42 ± 0.29 log₁₀ CFU/thigh. The bacterial densities in the 24-h-growth control animals are displayed in Fig. 2. In 7 of the 31 ESBL isolates tested, infection did not establish well, as evidenced by either a slight mean growth (≤0.31 log₁₀ CFU/thigh) or a decrease in log₁₀ CFU/thigh at 24 h despite the administration of saline instead of antibiotic. Therefore, these nonvirulent isolates were removed from the efficacy analysis. These isolates had a wide MIC range of 0.047 to 32 μg/ml for meropenem and 0.023 to 128 μg/ml for ertapenem. In general, a 1- to 4-log increase in CFU/thigh was observed at 24 h for the remaining 24 ESBL isolates (Fig. 2). The numbers of organisms recovered at 24 h ranged from 6.20 to 9.76 log₁₀ CFU/thigh, with a mean ± standard deviation of 7.87 ± 0.52 log₁₀ CFU/thigh. For isolates ATCC 25922 and E. coli 118, animals in the 24-h-growth control groups were harvested at 8 and 18 h, respectively, because the virulence of the established infection precluded their survival over the entire 24-h period. Animals infected with these organisms that received either meropenem or ertapenem survived to the completion of the study.

The abilities of meropenem and ertapenem to eradicate ESBL-producing E. coli and K. pneumoniae strains are displayed in Fig. 3. In general, both antibiotics display decreases in organisms at 24 h of between 1 and 3 log₁₀ for the majority of the isolates tested. For each ESBL isolate examined, encompassing a wide MIC range, meropenem displayed either decreased or static activity. Bacteriostasis (−0.04-log kill) was observed even for K. pneumoniae 262 (MIC = 8 μg/ml), which is considered intermediate according to CLSI-defined breakpoints. On the other hand, ertapenem resulted in bacterial kill up to an MIC of 1.5 μg/ml, above which bacteria grew between approximately 0.5 to 3.5 logs. Ertapenem, however, did show efficacy against K. pneumoniae 264, with an MIC of 4 μg/ml, which is defined as intermediate by CLSI susceptibility breakpoints.

No difference in bacterial kill was observed for either carbapenem for the majority of the ESBL isolates tested, because the MICs were low (≤0.5 μg/ml for meropenem; ≤1.5 μg/ml for ertapenem). This corresponded to fT>MICs of ≥75% and ≥23% for meropenem and ertapenem, respectively. For seven of the eight ESBL isolates with ertapenem MICs of ≥2 μg/ml, greater bacterial kill was observed in the mice given meropenem. This is most likely due to the two- to eightfold-greater in vitro potency of meropenem, which resulted in an fT>MIC of 30 to 65%, compared with the ertapenem fT>MIC of ≤20%.

For the two non-ESBL isolates tested, ATCC 25922 and E. coli 120, an approximate 2-log decrease in CFU/thigh was observed for both meropenem and ertapenem. Because of the low MICs of these organisms, the fT>MIC was high (77 to 93%). These isolates showed kills similar to those of ESBL isolates with similar MICs.

For the three isolates (E. coli 315, E. coli 321, and E. coli 322) that were tested at a 100-fold-higher inoculum (10⁷), the bacterial density for the 0-h control group was 7.40 ± 0.08 log₁₀ CFU/thigh. Unlike at the standard 10⁵ inoculum, the 24-h control group did not grow well, with a mean difference of −0.08 ± 0.33 log₁₀ CFU/thigh. Despite the higher MICs at this inoculum, which led to corresponding fT>MICs of 30 to 45% and 20 to 30% for meropenem and ertapenem, respectively, approximately 2- to 3.5-log decreases in bacteria were observed at 24 h. No differences in efficacy were observed between meropenem and ertapenem.