|Home | About | Journals | Submit | Contact Us | Français|
We have previously demonstrated that a high-dose, prolonged-infusion meropenem regimen (2 g every 8 h [q8h]; 3-hour infusion) can achieve 40% free drug concentration above the MIC against Pseudomonas aeruginosa with MICs of ≤16 μg/ml. The objective of this experiment was to compare the efficacy of this high-dose, prolonged-infusion regimen against carbapenemase-producing Klebsiella pneumoniae isolates with the efficacy against P. aeruginosa isolates having similar meropenem MICs. An in vitro pharmacodynamic model was used to simulate human serum concentrations. Eleven genotypically confirmed K. pneumoniae carbapenemase (KPC)-producing isolates and six clinical P. aeruginosa isolates were tested for 24 h, and time-kill curves were constructed. High-performance liquid chromatography (HPLC) was used to verify meropenem concentrations in each experiment. Meropenem achieved a rapid ≥3 log CFU reduction against all KPC isolates within 6 h, followed by regrowth in all but two isolates. The targeted %fT>MIC (percent time that free drug concentrations remain above the MIC) exposure was achieved against both of these KPC isolates (100% fT>MIC versus MIC = 2 μg/ml, 75% fT>MIC versus MIC = 8 μg/ml). Against KPC isolates with MICs of 8 and 16 μg/ml that did regrow, actual meropenem exposures were significantly lower than targeted due to rapid in vitro hydrolysis, whereby targeted %fT>MIC was reduced with each subsequent dosing. In contrast, a ≥3 log CFU reduction was maintained over 24 h for all Pseudomonas isolates with meropenem MICs of 8 and 16 μg/ml. Although KPC and P. aeruginosa isolates may share similar meropenem MICs, the differing resistance mechanisms produce discordant responses to a high-dose, prolonged infusion of meropenem. Thus, predicting the efficacy of an antimicrobial regimen based on MIC may not be a valid assumption for KPC-producing organisms.
Enteric Gram-negative bacteria are frequently responsible for serious nosocomial infections, such as bacteremia and pneumonia, in critically ill hospitalized patients. The increase in resistance among these bacteria, most notably Klebsiella spp. and Escherichia coli, due to the production of extended-spectrum β-lactamases (ESBLs) has led to the frequent use of carbapenem antibiotics. Carbapenems are considered by many to be the agents of choice for serious infections caused by ESBL-producing bacteria that may also display resistance to fluoroquinolones and aminoglycosides (23). Resistance to carbapenems among these bacteria remains remarkably rare in most countries. However, Klebsiella spp. that produce serine-based carbapenemase enzymes, referred to as Klebsiella pneumoniae carbapenemases (KPCs), have been identified in recent years (5).
The first KPC (KPC-1) was detected in a K. pneumoniae strain isolated in North Carolina in 2001 (28). Subsequently, KPC-2 and KPC-3 were identified from isolates from the mid-Atlantic and New England regions of the United States (1, 26, 27). Since the discovery of these initial KPC enzymes, numerous other variants (KPC-4 to KPC-7) have been reported internationally, as well as in other bacterial species, such as E. coli, Enterobacter spp., and Pseudomonas aeruginosa (9, 11, 13, 21, 25). Surveillance studies of the New York area have uncovered trends in the emergence of this resistance genotype (13). When comparing isolates gathered from the years 1999, 2001, and 2006, K. pneumoniae isolates have shown an alarming decline in antimicrobial susceptibility. The percentage of isolates that are resistant to carbapenems rose to 25%, and KPCs were detected in 38% of isolates. The number of multidrug-resistant K. pneumoniae isolates also increased to 59% in 2006, with 20% being resistant to all β-lactams, fluoroquinolones, and amikacin. These decreases in antimicrobial susceptibility have limited the number of therapeutic choices. In some studies, only polymyxins and tigecycline retain consistent susceptibility against KPC-producing Klebsiella spp. (2, 11).
While most KPC-producing K. pneumoniae isolates have carbapenem MICs above the breakpoint for susceptibility, some isolates can have imipenem and meropenem MICs as low as 1 to 2 μg/ml, and a number of them have meropenem MICs of 8 μg/ml and 16 μg/ml (5, 9, 21, 25). Standard dosages of meropenem (1 g every 8 h [q8h]) and imipenem (500 mg q6h) would not be expected to have clinical utility against organisms with MICs of 8 to 16 μg/ml.
Due to the lack of new antimicrobials with activity against multidrug-resistant, Gram-negative bacteria, thought leaders have suggested the application of pharmacodynamic principles to currently available antibiotics as our best opportunity to address these difficult-to-treat organisms (6). For β-lactams, a combination of administering higher doses and extending the infusion leads to increases in the time that free drug concentrations remain above the MIC (fT>MIC). In animal infection models, carbapenems require at least 40% fT>MIC to achieve maximal bactericidal activity against P. aeruginosa (6). A clinical pharmacodynamic study of patients with respiratory tract infections also identified a similar target (54% fT>MIC) as predictive of microbiological success (16). Higher dosages (2 g q8h) and prolonged infusions (3 h) have been applied to meropenem in numerous reports to improve the probability of achieving optimal bactericidal exposure against resistant Pseudomonas aeruginosa, Acinetobacter spp., and Burkholderia cepacia (7, 19) isolates. In pharmacokinetic simulation models, a 2-g q8h dose regimen of meropenem, with each dose infused over 3 h, has demonstrated a high probability of attaining 40% fT>MIC against organisms with meropenem MICs of up to 16 μg/ml (12). In theory, this dosage should maintain bactericidal activity against all organisms with MICs of up to 16 μg/ml, including KPC-producing isolates. Given the scarcity of new antibiotic options against Klebsiella spp. that harbor the KPC carbapenemase and its inevitable spread across hospitals as a cause of serious infections, the objective of this study was to assess the performance of a pharmacodynamically optimized meropenem regimen against KPC-producing Klebsiella isolates compared with P. aeruginosa isolates having the same meropenem MIC.
(This study was presented at the 49th Interscience Conference on Antimicrobial Agents and Chemotherapy, San Francisco, CA, 2009.)
Clinical K. pneumoniae isolates with KPC genotypes, all Hodge test positive and blaKPC positive (courtesy of Steve Jenkins, Mount Sinai Medical Center), and clinical P. aeruginosa isolates from our institution (Hartford Hospital) were screened for inclusion. Meropenem MICs for all isolates were determined by broth microdilution in triplicate according to Clinical Laboratory Standards Institute (CLSI) methodology, and the modal MIC was used for calculations in all experiments. Based on meropenem MICs, 11 K. pneumoniae isolates with the following MICs were selected: 2 μg/ml (n = 1); 8 μg/ml (n = 3); 16 μg/ml (n = 3); 32 μg/ml (n = 3); and 64 μg/ml (n = 1). Six P. aeruginosa isolates with MICs of 8 μg/ml (n = 3) and 16 μg/ml (n = 3) were included for comparison.
To confirm that the six P. aeruginosa isolates did not produce a carbapenemase, the modified Hodge test was applied as previously described (22). Briefly, the indicator organism, Escherichia coli ATCC 25922, at a turbidity of a 0.5 McFarland standard, was used to inoculate the surface of a Mueller-Hinton agar plate. After the plate was allowed to dry at room temperature for 3 to 10 min, a 10-μg meropenem susceptibility disk was placed in the center of the plate. Each of the six P. aeruginosa test isolates was then heavily streaked from the edge of the disk to the plate periphery on individual plates. Klebsiella pneumoniae ATCC BAA-1706 was used as the negative control, and Klebsiella pneumoniae ATCC BAA-1705 was used as the positive control. The plates were incubated overnight at 35°C in ambient air for 16 to 24 h. After incubation, the plates were examined for a cloverleaf type of indentation at the intersection of the test organism and the E. coli ATCC 25922 within the zone of inhibition of the carbapenem susceptibility disk.
Meropenem for intravenous injection (50 mg/ml; Merrem; AstraZeneca Pharmaceuticals) was obtained from the pharmacy at Hartford Hospital (lot PH0082; expiration date, February 2011; Novaplus).
The free drug concentrations for a meropenem dose of 2 g q8h administered as a 3-hour infusion over a 24-hour period (i.e., three doses) were simulated. Concentration-time profiles were based on median parameter estimates from a population pharmacokinetic analysis of meropenem in infected patients (15). Median estimates were entered into WinNonLin Professional version 5.0.1 (Pharsight Corporation, Mountain View, CA) to simulate steady-state exposure. Protein binding was applied by multiplying each concentration by 0.97 over the 24-hour period. The simulated peak concentration for each dose at 3, 11, and 19 h was 38.4 μg/ml, and the trough at 0, 8, 16, and 24 h was 3.04 μg/ml.
The in vitro pharmacodynamic model used in the study has been described previously (10). Each experiment consisted of three independent models (two experimental treatment models and one growth control model), which ran simultaneously for each isolate. The models were placed in a 37°C circulating water bath for optimal temperature control, and magnetic stir bars were used in each model to ensure adequate mixing of contents. Cation-adjusted Mueller-Hinton broth (CAMHB) (Becton, Dickinson and Company, Sparks, MD) was used as the bacterial growth medium in all in vitro model experiments. The volume of growth medium used for meropenem in each of the models was approximately 1,000 ml, based on the half-life and flow rate calculations. Three independent starting inoculums of 106 CFU/ml were set up from an overnight culture of the test isolate for all model experiments. To ensure that the bacteria were in logarithmic growth phase prior to antimicrobial exposure, antibiotic experiments were started 0.5 h after inoculation of bacteria into models.
In order to simulate the rising concentrations of a 3-hour infusion, meropenem was injected into the model in 30-min intervals from time zero to 3 h while the infusion pump was off. After 3 h, fresh CAMHB was continuously pumped into each of the models by a peristaltic pump (Masterflex L/S, model no. 7524-40; Cole-Parmer Instrument Company) at a rate that simulated the distribution half-life of meropenem obtained from the above-mentioned study. At 4.5 h, the rate of the peristaltic pump was slowed to simulate the terminal half-life of meropenem. At 8 and 16 h, the peristaltic pump was stopped, and the dosing process was repeated for the 24-h experiment.
To assess bacterial density over time, samples were obtained from each model and serially diluted in normal saline. Aliquots of each diluted sample were plated for quantitative culture. The volume of the aliquots used for bacterial counts was either 10 or 100 μl, depending on the dilution used. Trypticase soy agar plates (100-mm diameter) with 5% sheep blood were used for quantitative determinations. After 18 to 24 h of incubation at 37°C, the change in log10 CFU/ml over the 24-h interval was calculated, and time-kill curves were constructed by plotting log10 CFU/ml against time. The lower limit of detection for bacterial density was 101 CFU/ml. MIC testing in triplicate was repeated for select KPC isolates to determine whether an MIC change had occurred over the 24-h experiment.
Samples of CAMHB taken from each of the treatment models were assayed for meropenem at 0, 0.5, 1, 1.5, 2, 2.5, 3, 4.5, 6, 7, 8, 9.5, 11, 14, 16, 17.5, 19, 22, and 24 h. All samples were immediately stored at −80°C until analyses. Samples were analyzed by a high-performance liquid chromatography (HPLC) method as described previously (8), validated for broth. The meropenem HPLC assay was linear (r > 0.997) over a concentration range from 0.25 to 40 μg/ml. The intraday quality control samples (n = 10 each) of 0.5 and 30 μg/ml had a percentage coefficient of variation (%CV) of 3.46 and 4.85, respectively. The interday quality control samples (n = 26) of 0.5 and 30 μg/ml had a %CV of 3.09 and 3.16, respectively.
Prior to bacterial experiments (i.e., in bacterium-free models), we developed a specific dosing administration strategy that consistently resulted in the targeted meropenem concentration-time profile depicted in Fig. Fig.1.1. However, in infected models, meropenem concentration-time profiles differed depending on the baseline MIC and isolate being tested. This resulted in achieving intended fT>MIC exposures for all P. aeruginosa isolates but only 3 of the 11 KPC-producing K. pneumoniae isolates (Table (Table1).1). The overall desired %fT>MIC of ≥40% was achieved for each dosing interval only in the K. pneumoniae isolate with an MIC of 2 μg/ml and two of the three isolates with a meropenem MIC of 8 μg/ml. Rapid in vitro hydrolysis of the meropenem β-lactam ring was presumed to be the cause for lower-than-intended meropenem concentrations, because in many of the experiments, no meropenem peak was noted on HPLC after dosing to simulate the second and third intervals. Importantly, while targeted meropenem fT>MIC exposures during the first 8-h dose interval were achieved against all KPC isolates with MICs of 8 μg/ml, drug exposure fell short of the target during the second and third 8-h dose intervals. A similar decline with each subsequent dosing was also noted for two of the three KPC isolates with MICs of 16 μg/ml. As a result, we were unable to achieve our targeted exposure of 47% fT>MIC during any of the experiments for these isolates.
The average bacterial density of the starting inoculum was (1.96 ± 0.47) × 106 CFU/ml and (7.54 ± 0.52) × 105 CFU/ml for K. pneumoniae and P. aeruginosa isolates, respectively. KPC control isolates grew to (3.06 ± 1.18) × 108 at 24 h, while P. aeruginosa control isolates grew to (3.29 ± 2.35) × 107. Figure Figure22 summarizes the time-kill curves for K. pneumoniae isolates, defined by baseline meropenem MIC. Meropenem achieved a rapid >3 log CFU reduction within 6 h against all K. pneumoniae isolates, followed by regrowth in all but two isolates. Targeted fT>MIC exposure was achieved during each of the 8-h dosing intervals for these isolates (KPC isolate 377, MIC = 2 μg/ml; KPC isolate 328A, MIC = 8 μg/ml). Although a fT>MIC exposure greater than 40% was achieved for each of the dosing intervals against KPC isolate 351 (MIC, 8 μg/ml), this isolate regrew back to control levels. In contrast, meropenem was observed to have rapid and sustained bactericidal activity (≥3 log reduction) against all P. aeruginosa isolates tested (MIC, 8 μg/ml and 16 μg/ml) over the 24-h experiment. Figures 3a and b summarize the resultant time-kill curves against P. aeruginosa.
Five of the 11 KPC isolates were selected to have the MIC for the organisms that regrew after exposure to meropenem retested. These isolates had baseline MICs of 8 μg/ml, 16 μg/ml (n = 2), and 32 μg/ml (n = 2). All isolates demonstrated increased meropenem MICs to >32 μg/ml after 24 hours of exposure to meropenem.
Carbapenems have historically been considered the agents of choice to treat infections caused by multidrug-resistant Enterobacteriaceae, but the increasing prevalence of isolates producing KPC carbapenemase has compromised the effectiveness of this class (21). Therapeutic options for infections caused by KPC-producing Klebsiella are scarce, with only the polymyxin antibiotics and tigecycline demonstrating consistent susceptibility. However, these antibiotics are not without their own limitations, thus additional treatment options are still needed for the successful treatment of these organisms. Since a number of KPC-producing K. pneumoniae isolates can have meropenem MICs in the susceptible (≤4 μg/ml), intermediate (8 μg/ml), or low-level resistant (16 μg/ml) ranges, this study sought to determine whether a human simulated, high-dose, prolonged-infusion meropenem regimen could produce and maintain a bactericidal effect comparable with that of P. aeruginosa having the same meropenem MICs. The free drug exposure from a meropenem dose of 2 g q8h, with each dose infused over 3 h, was selected for this study because this dosing regimen has been used clinically (7, 20) and is not greater than the maximum approved daily dose recommended by the package insert, which has demonstrated safety and tolerability, and importantly, because previous Monte Carlo simulation studies have demonstrated that this dosing regimen has a high probability of achieving at least 40% fT>MIC at MICs of up to 16 μg/ml, which is the exposure that produced maximal bactericidal activity against P. aeruginosa in murine infection models (12, 19).
Despite administering meropenem in dosages consistent with achieving the targeted exposure for each isolate, this optimized regimen failed to maintain bactericidal activity against 9 of the 11 studied KPC-producing K. pneumoniae isolates. Although rapid bactericidal exposure was observed over the first 6 h against all KPC isolates, all but two (MICs of 2 μg/ml and 8 μg/ml) regrew back to control isolate levels by 12 to 16 h. The simulated meropenem dosing regimen aimed at achieving 100% fT>MIC against the organism with an MIC of 2 μg/ml. This exposure was accomplished, resulting in bactericidal activity within 1 h of dosing that was retained throughout the 24-h period (Table (Table1).1). The second isolate that responded to this meropenem regimen had an MIC of 8 μg/ml; 69% fT>MIC was targeted and attained against this isolate, resulting in a bactericidal response that, albeit slower, was maintained over the 24-h experiment (Table (Table11).
Unexpectedly, the remaining two KPC isolates having a meropenem MIC of 8 μg/ml regrew to control levels (Fig. (Fig.2).2). Moreover, a substantial loss of meropenem was observed in the models as analyzed by HPLC, likely due to hydrolysis of the carbapenem structure. While the targeted exposure (69% fT>MIC) was achieved against all isolates with an MIC of 8 μg/ml during the first 8-h dosing interval, meropenem exposure began to decline with subsequent doses for KPC isolates 351 and 354. Although one of these isolates (KPC isolate 351) was still exposed to meropenem concentrations that remained above the MIC for 40% fT>MIC, this organism regrew. KPC isolate 354 (MIC of 8 μg/ml) was exposed to meropenem concentrations for 56% and 3% of the dosing interval after the second and third dosages, respectively. Not surprisingly, this isolate also regrew to control levels and had a 24-h repeat MIC that was now >32 μg/ml. A similar observation was noted for the three isolates with meropenem MICs of 16 μg/ml. Targeted exposure was 47% fT>MIC, which was not achieved during any of the dosing intervals, with the majority of the exposures being 0% fT>MIC. All three of these isolates regrew to control levels at 24 h, and repeat MICs increased to >32 μg/ml in the two isolates that were retested. As expected, all KPC isolates with baseline meropenem MICs of 32 μg/ml and 64 μg/ml regrew to control levels, as targeted exposure was 16% fT>MIC and 0% fT>MIC, respectively. However, despite administering the same amount of meropenem to the models as in all other experiments, detectable meropenem concentrations above the MIC for these organisms were not observed. Because of these observations, these experiments were repeated independently several times on different days, with consistent findings in each.
Antimicrobial regimens are usually driven by the phenotypic profile rather than the genotypic profile of the organisms. Previous work by our group with a murine thigh infection model has demonstrated that regardless of the presence of an ESBL, which demonstrated a significant in vitro inoculum effect on cefepime MIC, the desired bactericidal effect was still achieved as long as the cefepime exposure was ≥40% fT>MIC (18). This would suggest that in organisms with increased meropenem MICs, resistance mechanisms, such as the production of β-lactamases, can be overcome with the same fT>MIC exposure as that required for wild-type isolates. Clearly, this was not the case in the current study against the KPC-producing K. pneumoniae isolates, where two of the three isolates at 8 μg/ml regrew to control levels despite achieving >40% fT>MIC for the first and second dosing intervals. Our observations may reflect an in vitro/in vivo paradox, since previous work conducted by Craig and colleagues (3) suggested that the presence of KPCs in Enterobacteriaceae had no impact on the T>MIC required to produce bacteriostasis with doripenem, meropenem, or imipenem in the murine infection model; however, many of those isolates had much lower meropenem MICs than those of the clinical isolates from our study. Similarly, we observed bacteriostatic effects when examining a human simulated doripenem dose of 1 g q8h administered as a 4-h infusion against these same clinical isolates (doripenem MICs, 4 to 32 μg/ml) in the murine thigh infection model. In contrast, this doripenem regimen has previously demonstrated consistent 2 to 3 log reductions in P. aeruginosa CFU with doripenem MICs of up to 8 μg/ml, and some variable killing against isolates with MICs as high as 16 μg/ml (4).
Against the six clinical P. aeruginosa isolates with MICs of 8 and 16 μg/ml, the simulated meropenem regimen achieved a >3 log reduction in CFU over the first 8 to 12 h, with minimal regrowth of the isolates by 24 h (Fig. (Fig.3).3). Furthermore, free meropenem concentrations remained above the MIC for at least the targeted 69% and 47% of the dosing interval, respectively. Importantly, isolates from these experiments did begin to regrow by 24 h, but colony counts were still well below that of the control, and were significantly different from the KPC isolates at MICs of 8 and 16 μg/ml.
Nevertheless, it is notable that the killing profile of this simulated meropenem regimen differed between the KPC and P. aeruginosa isolates even though the baseline MICs were similar. This divergence may be in part due to the differing resistance mechanisms of these species. The predominant mechanisms of carbapenem resistance within P. aeruginosa isolates arise from the interaction between efflux systems, ampC expression, and loss of outer membrane porins (deletions in oprD expression) (17, 24). Furthermore, we demonstrated that the six P. aeruginosa isolates tested did not produce a carbapenemase enzyme according to the modified Hodge test. In contrast, meropenem resistance among these K. pneumoniae isolates is due predominantly to the production of the serine carbapenemase. Importantly, other resistance mechanisms, such as the production of additional β-lactamases (i.e., ESBLs), ampC expression, and loss of OmpK36, can contribute to carbapenem resistance in Klebsiella spp. (14). A limitation to our experiment is the absence of data about detailed resistance mechanisms of these Klebsiella isolates other than their known expression of blaKPC. It has been demonstrated that many of the clinical KPC isolates from New York City produce several β-lactamases, in addition to the serine carbapenemase (9). These additional resistance mechanisms may have been responsible for the discordance in meropenem effect among the three KPC isolates with an MIC of 8 μg/ml. Further work is required to determine which resistance mechanisms may have contributed to increasing meropenem MICs among the tested KPC isolates during the experiment.
In conclusion, carbapenemase-producing K. pneumoniae can have a meropenem MIC similar to that of P. aeruginosa but exhibits markedly different bactericidal responses to a pharmacodynamically optimized regimen, due to the different genotypic resistance mechanisms. Despite the proposed effectiveness of the optimized meropenem regimen against organisms with MICs of up to 16 μg/ml, a reliable reduction in the bacterial density is not seen in the presence of serine-based carbapenemases. While the phenotypic profile of an organism has historically driven the efficacy of the antimicrobial regimen, these assumptions may no longer be valid in the case of carbapenemase-producing K. pneumoniae isolates.
This study was funded by a research grant from AstraZeneca Pharmaceuticals, Inc., Wilmington, DE. D.P.N. and J.L.K. have received research support and consultancy support and are members of the speakers' bureau for AstraZeneca.
We acknowledge Stephen Jenkins for kindly providing KPC-positive K. pneumoniae isolates.
Published ahead of print on 7 December 2009.