Optimizing Doses of Ceftolozane/Tazobactam as Monotherapy or in Combination with Amikacin to Treat Carbapenem-Resistant Pseudomonas aeruginosa

Carbapenem-resistant Pseudomonas aeruginosa (CRPA) is a hospital-acquired pathogen with a high mortality rate and limited treatment options. We investigated the activity of ceftolozane/tazobactam (C/T) and its synergistic effects with amikacin to extend the range of optimal therapeutic choices with appropriate doses. The E-test method is used to determine in vitro activity. The optimal dosing regimens to achieve a probability of target attainment (PTA) and a cumulative fraction of response (CFR) of ≥90% were simulated using the Monte Carlo method. Of the 66 CRPA isolates, the rate of susceptibility to C/T was 86.36%, with an MIC50 and an MIC90 of 0.75 and 24 µg/mL, respectively. Synergistic and additive effects between C/T and amikacin were observed in 24 (40%) and 18 (30%) of 60 CRPA isolates, respectively. The extended infusion of C/T regimens achieved a ≥90% PTA of 75% and a 100% fT > MIC at C/T MICs of 4 and 2 µg/mL, respectively. Only the combination of either a short or prolonged C/T infusion with a loading dose of amikacin of 20–25 mg/kg, followed by 15–20 mg/kg/day amikacin dosage, achieved ≥90% CFR. The C/T infusion, combined with currently recommended amikacin dose regimens, should be considered to manage CRPA infections.


Introduction
Pseudomonas aeruginosa is a significant cause of hospital-acquired infections and is frequently multidrug-resistant (MDR) [1]. MDR P. aeruginosa was reported in 14.7% and 22.0% of cases of bloodstream infections and pneumonia, respectively. The high mortality rate of infections with MDR P. aeruginosa makes treating serious infections more challenging. Furthermore, MDR P. aeruginosa is frequently resistant to carbapenems and other β-lactams. β-lactam resistance is mediated via multiple mechanisms, including the acquisition of metallo-β-lactamases, the increased production of chromosomal AmpC, increased efflux, and changes in membrane permeability [2,3]. The most consistently active drugs against MDR P. aeruginosa are aminoglycosides and polymyxins, but pharmacokinetic limitations and their association with worse outcomes when given as monotherapy make their use suboptimal [4,5].
C/T was the most active β-lactam agent against the 66 strains of CRPA from the E-test method, of which 86.36% showed susceptibility at an MIC 50 of 0.75 µg/mL and an MIC 90 of 24 µg/mL.

PTA and CFR
The PTA, including 75% and 100% f T > MIC for ceftolozane and 20% f T ≥ 1 µg/mL for tazobactam, for each dosing regimen at specific MICs is shown in Tables 4 and 5. All antibiotic dosing regimens (detailed in Section 4.4) met the criteria of ≥90% PTA of 20% f T ≥ 1 µg/mL for tazobactam. The PTA of achieving 75% f T > MIC following ceftolozane administration by a 0.5-h infusion, a prolonged 4-h infusion, and a loading dose followed by continuous infusion was 93.92%, 93.61%, and 93.61% at MICs of 2, 4, and 8 µg/mL, respectively. Furthermore, the PTA of achieving 100% f T > MIC following ceftolozane administration by a 0.5-h infusion, a prolonged 4-h infusion, and a loading dose followed by continuous infusion was 95.50%, 94.75%, and 93.60% at MICs of 1, 2, and 8 µg/mL, respectively. None of the C/T dosing regimens achieved the PTA target when the MIC was ≥16 µg/mL.  For C/T, only combination regimens achieved the target of ≥90% CFR at 75% and 100% f T > MIC. None of the amikacin dosing regimens achieved the CFR target when administered as monotherapy. Fortunately, when amikacin was combined with C/T, the amikacin dosage, a loading dose of 20-25 mg/kg, followed by 15-20 mg/kg every 24 h, met the CFR target. The CFR for each dosing regimen of C/T, amikacin, and its combination are presented in Tables 6 and 7. Table 6. The cumulative fraction of response (CFR) for each dosing regimen of ceftolozane/tazobactam (C/T) and ceftolozane/tazobactam (C/T), combined with amikacin (AMK).

Discussion
There are limited treatment options for CRPA infections. The agents to which CRPA is susceptible include colistin or polymyxin B, which exhibit a suboptimal treatment outcome and a high rate of adverse reactions, especially nephrotoxicity. Our study in a Thai university hospital demonstrated the attractive in vitro activity of C/T against CRPA, with a susceptibility rate of 86.36%, which correlated with other studies, where the susceptibility rate of MDR P. aeruginosa or CRPA ranged from 67.2-85.9% [9,16]. Although C/T is considered the most active β-lactam against both susceptible and resistant strains of P. aeruginosa, a study from Singapore revealed a susceptibility rate of 37.9% for CRPA to C/T. This discrepancy was mainly attributed to the presence of carbapenemase-producing isolates, particularly metallo-β-lactamases [17].
Due to the high rate of susceptibility of CRPA to C/T, C/T may be useful as an empirical or definite antibiotic therapy for the treatment of infections suspected or known to be caused by CRPA. A clinical study showed that C/T was successful in treating 71% of patients with MDR P. aeruginosa infections [18]. Therefore, a polymyxin-sparing strategy may reduce antibiotic-related adverse events, as well as minimize the over-prescription of polymyxin.
Among the few reports of the synergistic effect of C/T with other antibiotics, studies involving isolates of P. aeruginosa are limited. Antipseudomonal antibiotics that produce a synergistic effect with C/T include aztreonam, amikacin, and meropenem [10][11][12]. In our study, we observed 40% synergism between C/T and amikacin against CRPA isolates using E-test methods. A recent study from Greece performed synergistic testing between C/T and amikacin against MDR P. aeruginosa using a time-kill assay, and a synergistic effect was observed in 85% [19]. These differences and discordant results may be due to several factors. First, the agreement between the time-kill assay and E-test crossing method ranged from 3-71% in MDR Gram-negative bacilli, including P. aeruginosa [20]. Second, differences in the genotypic resistance of bacterial strains may affect the synergistic result [19]. The good bactericidal and synergistic activity observed with the combination of C/T and amikacin may be attributed to β-lactam-mediated membrane damage, leading to increased aminoglycoside uptake [21]. Clinical data regarding the treatment outcome of C/T combination therapy against MDR Gram-negative bacilli, mostly MDR P. aeruginosa, revealed a significant decrease in mortality. A systematic review and meta-analysis included 8 non-randomized studies of C/T for treatment as monotherapy and combination therapy. The results showed that C/T in combination was associated with clinical improvement (OR, 0.97; 95% CI, 0.54 to 1.74; p = 0.954) and statistically lower mortality at 30 days (OR, 0.31; 95% CI, 0.10 to 0.97; p = 0.045) than the patient receiving C/T monotherapy [22]. Furthermore, a successful treatment outcome and rapid microbiological clearance using combination therapy of C/T with tobramycin against C/T-resistant P. aeruginosa in a severely neutropenic patient were also reported [23]. Thus, C/T combination therapy is potentially beneficial when combating refractory infections of MDR P. aeruginosa. Aminoglycosides, particularly amikacin, should be considered as the combination agent with C/T to achieve a good synergistic or additive effect.
The magnitude of the PK/PD target for cephalosporins that provided a maximal bactericidal effect was reported to range from 60-70% f T > MIC in preclinical studies, whereas the magnitude of the cephalosporin PK/PD target to achieve a clinical cure and a microbiological cure was reported as 100% and 60-100% f T > MIC, respectively, in clinical studies [24]. When using ceftolozane against P. aeruginosa, the % f T > MIC to achieve a 1-or 2-log reduction ranged from 30-40% [25,26]. Therefore, using 75% and 100% f T > MIC as the PK/PD targets of ceftolozane in the simulated regimens may be adequate to predict the maximum bactericidal effect and microbiological cure.
The simulation studies revealed that antibiotic dosing regimens by prolonged infusion or continuous infusion had greater potential than intermittent infusion. Prolonged infusion or continuous infusion regimens achieved a target of ≥90% PTA of 75% and 100% f T > MIC with C/T MICs ≤ 4 µg/mL (CLSI susceptible breakpoint) and ≤2 µg/mL, respectively, whereas a 1.5 g intermittent infusion every 8 h (an approved C/T dosing regimen for complicated urinary tract and intraabdominal infections) achieved ≥90% PTA at C/T MICs of ≤2 and ≤1 µg/mL, respectively, which agreed with previous studies [27,28]. There are concerns about the stability of C/T when using prolonged infusion or continuous infusion regimens. However, it was recently reported that C/T is stable for at least 24 h in 0.9% normal saline and 5% glucose solution in real-world conditions when stored in polypropylene tubes at room temperature (22 • C) without light exposure [29]. Therefore, the extended infusion of C/T is feasible, and because it increases the probability of treatment success, it may be a recommended regimen for the treatment of infections.
The synergistic effects of C/T plus amikacin contributed to achieving a target of ≥90% CFR. None of the prolonged or continuous infusions of C/T monotherapy achieved the target CFR. A previous study showed that only a continuous infusion of C/T monotherapy met the target of ≥85% CFR at 40%, 60%, and 100% f T > MIC [28]. However, when C/T combined with amikacin, all C/T dosing regimens except intermittent regimens and all amikacin dosing regimens (loading dose 20-25 mg/kg, followed by 15-20 mg/kg/day) reached the CFR target of ≥90%. In 2017, Kato et al. recommended that the amikacin initial dose required to achieve C max /MIC ≥ 8 was 15 mg/kg/day, and the amikacin maintenance dosage was 15 mg/kg/day at amikacin MICs ≤ 4 µg/mL [30]. Fortunately, all clinical CRPA isolates except no.47 had amikacin MICs ≤ 4 µg/mL when combined with C/T. If synergism occurs, the MIC values of each antibiotic were reduced by at least 1-fold dilution. The decrease in MIC affects an increase f T > MIC for C/T and C max /MIC for amikacin, respectively, resulting in a greater achievement of the probability of CFR target in each antibiotic. Thus, it may be advantageous to consider C/T plus amikacin as an empirical therapy.
To our knowledge, this is the first study to determine the in vitro susceptibility and synergistic effect of C/T against CRPA isolates in Thailand. However, several limitations were encountered. First, the BMD method, which is the gold standard of antimicrobial susceptibility testing, was not performed for C/T. However, the E-test method for C/T is a simple and acceptable method for susceptibility testing [31]. Second, the E-test crossing method was selected as the synergistic testing method in this study. It is widely used in clinical practice because it is easy to perform. However, the other methods for synergistic study, especially time-kill assay as the gold standard, can be evaluated. Third, the genotypic resistance characteristics of the CRPA isolates were not investigated. Thus, the molecular basis of the characteristics of these CRPA strains, which might contribute to a better understanding the results, were not explored. Fourth, this was a limited single-center study, which may affect the generalizability of the CFR results. Thus, our findings should be appraised and compared with other cohorts. Furthermore, we recommend that a nationwide multicenter study using standardized antimicrobial susceptibility testing methods based on various types of CRPA should be undertaken. Fifth, the antibiotic dosing regimens were simulated using typical pharmacokinetic parameters [28]. Thus, antibiotic dosing regimens for C/T and for amikacin based on creatinine clearance should be determined and monitored by the therapeutic drug monitoring (TDM) for amikacin in order to be more effective and less nephrotoxic. Finally, our clinical CRPA isolates were mostly susceptible to amikacin; thus, the optimal amikacin dosing recommendation in the combination therapy should be used if any CRPA isolates are susceptible to amikacin. Despite these limitations, this study provides essential information for treating P. aeruginosa in clinical practice, particularly CRPA, where treatment options are extremely limited. In summary, C/T had a fair synergistic effect with amikacin and may be considered as a combination therapy in CRPA infection.

Bacterial Identification and Antimicrobial Susceptibility Test
From January-December 2020, CRPA isolates were collected from patients by the microbiology laboratory at Phramongkutklao Hospital, a 1200-bed teaching hospital of the Phramongkutklao College of Medicine, Royal Thai Army, Bangkok, Thailand. All studied CRPA isolates were identified as P. aeruginosa using matrix-assisted laser desorption ionization time-of-flight mass spectrometry.
The clinical isolates of CRPA were selected and underwent antimicrobial susceptibility testing against C/T and amikacin. The MIC values of each studied antibiotic were determined using E-test strips (Liofilchem, Inc., Waltham, MA, USA). A purified colony of isolated strains was picked up and suspended to 0.9% in normal saline (Univar ® , Ajax Finechem Pty Ltd., Taren Point, Australia) as 0.5 McFarland standard. The susceptibility data of C/T and amikacin against the CRPA isolates were collected and analyzed. The MICs of C/T and amikacin ranged from 0.008/4-128/4 µg/mL to 0.016-256 µg/mL, respectively. The results were interpreted according to the criteria of the Clinical and Laboratory Standards Institute (CLSI) [32]. All CRPA strains were stored at −80 • C until analysis. P. aeruginosa ATCC 27853 was used as a reference strain for quality control.
MIC values, MIC 50 , and MIC 90 were measured for a tested population. MIC 50 and MIC 90 are defined as the MIC values inhibiting 50% and 90% of the tested isolates, respectively.

Synergy Test of C/T against CRPA
We designed a synergy test for C/T with amikacin against CRPA. The MIC values of each studied antibiotic were initially determined in order to further perform synergistic testing. For the synergy test, two E-test strips of studied antibiotics crossing formation in each MIC value with 90 • angle were placed on an inoculated Mueller-Hinton Agar plate with bacteria spread. The resulting ellipse of inhibition was checked after 16-18 h at 35 ± 2 • C.
The fractional inhibitory concentration index (FICI) was calculated for each antibiotic in each combination using the following formula: FICA + FICB = ∑FICI, where FICA equals the MIC of drug A in combination divided by the MIC of drug A alone, and FICB equals the MIC of drug B in combination divided by the MIC of drug B alone. The ∑FICI were interpreted as follows: synergy, FICI ≤ 0.5; additive effect, FICI > 0.5-≤ 1; no interaction (indifference), FICI > 1-≤ 4; or antagonism, FICI > 4 [33,34].

Phenotypic Classification
The CRPA isolates were classified as MDR (resistant to at least one agent in three or more antimicrobial categories), XDR (resistant to at least one agent in all but two or fewer antimicrobial categories), or PDR (resistant to all agents in all antimicrobial categories) based on the CLSI criteria described by Magiorakos et al. [35].

Antibiotic Dosing Regimen Simulations
We used two-compartment pharmacokinetic models of C/T and amikacin with linear elimination to simulate the plasma concentration time [28,36]. Pharmacokinetic parameters, the PK/PD targets, and indices for the simulation are described in Table 8. Table 8. Pharmacokinetic parameters, the PK/PD targets, and indices of ceftolozane/tazobactam (C/T) and amikacin used for the simulation. Simulated dosing regimens of C/T with log-normal distributions included intermittent infusion (1.5 g infusion over 0.5 h every 8 h) and extended infusion (1.5 g infusion over 4 h every 8 h and 1.5 g loading dose over 0.5 h, followed by 4.5 g continuous infusion over 24 h). The simulated dosing regimens of amikacin with log-normal distributions included a loading dose of 20-25 mg/kg, followed by maintenance doses of 15-20 mg/kg every 24 h.

Antibiotics
The optimal antibiotic dosing regimens were simulated using a 10,000-subject Monte Carlo simulation (Oracle Crystal Ball Classroom Faculty Edition-Oracle 1-Click Crystal Ball 201, Thailand). PK/PD targets were set for each studied antimicrobial. The percentages of targets for the duration of time (f T) that the free drug concentration had to remain above the MIC (f T > MIC) were 75% and 100% for ceftolozane, and the target for tazobactam was 20% f T ≥ 1 µg/mL. The target ratio between the maximum drug concentration obtained after a single dose and the MIC (C max /MIC) was ≥ 8 for amikacin.
The probability of target attainment (PTA) was determined as the percentage of probability at which the pharmacodynamic indices with specific MICs were achieved. The cumulative fraction of response (CFR) was calculated as the proportion of % PTA for each MIC of each of the pharmacodynamic indices according to the MIC distribution [37]. At least 90% PTA at a steady state for documented therapy and 90% CFR were considered the achievement targets for empirical therapy.

Conclusions
The CRPA clinical isolates from Thailand in our study showed high in vitro susceptibility (86.36%) to C/T. Furthermore, a 40% synergistic effect was observed in combination with amikacin. C/T extended infusion regimens may be considered empirical or definite antibiotic therapies when CRPA is suspected or detected. Amikacin with a loading dose of 20 mg/kg, followed by 15 mg/kg/day, seems an attractive combination agent with C/T when combination therapy is necessary.