Effect of ceftazidime-avibactam combined with different antimicrobials against carbapenem-resistant Klebsiella pneumoniae

ABSTRACT This study aimed to assess the in vitro efficacy of ceftazidime-avibactam (CZA) in combination with various antimicrobial agents against carbapenem-resistant Klebsiella pneumoniae (CRKP). We selected 59 clinical CRKP isolates containing distinct drug resistance mechanisms. The minimum inhibitory concentrations (MICs) of meropenem (MEM), colistin (COL), eravacycline (ERA), amikacin (AK), fosfomycin (FOS), and aztreonam (ATM), both individually and in combination with CZA, were tested using the checkerboard method. The interactions of antimicrobial agent combinations were assessed by fractional inhibitory concentration index (FICI) and susceptible breakpoint index (SBPI). The time-kill curve assay was employed to dynamically evaluate the effects of these drugs alone and in combination format. In the checkerboard assay, the combination of CZA+MEM showed the highest level of synergistic effect against both KPC-producing and carbapenemase-non-producing isolates, with synergy rates of 91.3% and 100%, respectively. Following closely was the combination of FOS+CZA . For metallo-beta-lactamases (MBLs) producing strains, ATM+CZA displayed complete synergy, while the combination of MEM+CZA showed a synergy rate of only 57.14% for NDM-producing strains and 91.67% for IMP-producing strains. In the time-kill assay, MEM+CZA also demonstrated significant synergistic effects against the two KPC-2-producing isolates (Y070 and L70), the two carbapenemase-non-producing isolates (Y083 and L093), and the NDM-1-producing strain L13, with reductions in log10 CFU/mL exceeding 10 compared to the control. Against the IMP-producing strain Y047, ATM+CZA exhibited the highest synergistic effect, resulting in a log10 CFU/mL reduction of 10.43 compared to the control. The combination of CZA and MEM exhibited good synergistic effects against KPC-producing and non-enzyme-producing strains, followed by the FOS+CZA combination. Among MBL-producing strains, ATM+CZA demonstrated the most pronounced synergistic effect. However, the combinations of CZA with ERA, AK, and COL show irrelevant effects against the tested clinical isolates. IMPORTANCE Our study confirmed the efficacy of the combination CZA+MEM against KPC-producing and non-carbapenemase-producing strains. For metalloenzyme-producing strains, CZA+ATM demonstrated the most significant synergy. Additionally, CZA exhibited a notable synergy effect when combined with FOS. These combination therapies present promising new options for the treatment of CRKP infection.

Ceftazidime-avibactam (CZA) is a novel combination of a third-generation cepha losporin and a novel β-lactamase inhibitor (6), demonstrating efficacy against CRKP.Although CZA is currently a significant drug for the treatment of CRKP, several researches have elucidated the emergence of resistance to CZA (7).Some investigations showed the emergence of CZA resistance due to a single alternative mutation in the KPC-2 or KPC-3 gene (8,9).With the increasing use of this drug, the prevalence of resistance is expected to escalate, rendering monotherapy less effective for severe infections.Combination therapy has demonstrated effectiveness against a broad spectrum of resistant organisms, ensuring maximum antimicrobial efficacy (9).This approach not only diminishes the reliance on single agents but also mitigate the emergence of drug resistance.Notably, synergistic effects between different antibacterial drugs have been observed, further emphasizing the potential benefits of combination strategies.
This study aimed to investigate the efficacy of CZA, both as a standalone treatment and in combination with other clinically recommended drugs, including meropenem (MEM), colistin (COL), eravacycline (ERA), amikacin (AK), fosfomycin (FOS), and aztreonam (ATM).Notably, the combination of ATM and CZA was exclusively assessed in strains producing MBL.
All strains were identified using the MALDI-TOF MS apparatus (Bruker Biotyper; Bruker Daltonik, Bremen, Germany).Mueller-Hinton broth and Luria-Bertani agar broth were used for susceptibility testing and time-kill experiments, respectively.

MLST and carbapenemase detection
Multilocus sequence typing (10) (MLST) was performed as described on the Pasteur Institute MLST website (https://bigsdb.pasteur.fr/),including DNA sequencing analysis of the seven housekeeping genes.The carbapenemase gene (11) (multiplex) probe-based Hi-Media Hi-PCR kit was used to detect specific regions of the gene encoding the carbapenemase enzymes (Table S1).The phenotypic Carbapenem Inactivation Method (CIM) was conducted following the CLSI recommended guidelines to examine whether strains could produce carbapenemase.Briefly, 1 µL of overnight-cultured bacteria was mixed with 2 mL of tryptic soy broth (TSB), and 10 mg of meropenem disk was added; incubated the suspension at 37°C for 4 h.After incubation, the disk was removed from the suspension using an inoculation loop, placed on a Mueller-Hinton agar plate inoculated with a susceptible Escherichia coli indicator strain (ATCC 29522) and subse quently incubated at 35°C.Disks incubated in suspensions that do not contain carbape nemases would yield a clear inhibition zone (12).
The results were interpreted according to CLSI criteria.E. coli strain ATCC 25922 and Pseudomonas aeruginosa ATCC 27853 were used as quality control organisms.The MIC results were measured after 16-20 h of incubation in the air at 35℃ and interpreted as susceptible, intermediate, or resistant according to CLSI guidelines.

Checkerboard test
We used CZA as the basic antibacterial agent in combination with other drugs.Six combinations were tested including CZA+MEM, CZA+AK, CZA+COL, CZA+ERA, CZA+ATM and CZA+FOS.Taking the CZA+MEM combination as an example to dem onstrate the preparation of the drug susceptibility test plates, the lateral direction represents the MEM gradient dilution direction, while the longitudinal direction represents the CZA gradient dilution direction.The prepared starting solution is subjected to a series of two-fold dilution, and 25 µL of the solution was added to each well, forming drug susceptibility test plates with the concentration range of CZA from 0.031 μg/mL to 4 μg/mL and MEM from 0.5 μg/mL to 64 μg/mL.These plates were then stored at −80℃ for standby.The test strains were inoculated into the drug susceptibility test plates at a final concentration of 5 × 10 5 CFU/mL, then incubated the plates for 20 hours at 35℃ to judge MIC.
The efficacy of the antimicrobial combination was evaluated using the fractional inhibitory concentration index (FICI).The FICI was calculated using the following formula: FICI = (MIC of agent A in combination/MIC of agent A alone) + (MIC of agent B in combination/MIC of agent B alone) (13).The results were interpreted as follows: FICI ≤0.5 indicated synergistic, 0.5 ＜ FICI ≤4 implied irrelevant, FICI >4 is antagonistic (14).
Additionally, we employed the susceptible breakpoint index (SPBI) to evaluate the interaction of antimicrobial agents.SPBI was calculated as follows: SBPI ＝ (susceptible breakpoint A/MIC of A in combination) + (susceptible breakpoint B/MIC of B in combina tion) (15).

Time-kill assay
In order to dynamically assess the bactericidal activity of the antimicrobials, we chose six representative strains to conduct time-kill assay, which showed the synergistic effect for most of the six drug combinations.Detailed information about these six strains is presented in Table 1.The drug concentrations were determined based on reported blood concentrations that could be achieved for each drug.The individual drug concentrations, a MEM, meropenem; ERA, eravacycline; FOS, fosfomycin; COL, colistin; AK, amikacin; ATM, aztreonam; CZA, ceftazidime-avibactam;R,resistant; S,Sensitive.
b "-" indicates that the drug tested in KPC-producing as well as non-carbapenemase-producing strains did not include the AZA.
both in isolation and in combination, are as follows: MEM 8 µg/mL, COL 0.25 µg/mL, ERA 0.25 µg/mL, AK 16 µg/mL, FOS 32 µg/mL, ATM 1 µg/mL, CZA 0.5/4 µg/mL.Firstly, prepare 0.5 McFarland Standard Turbidity of colonies growing in the logarith mic growth overnight, and dilute with MHBII by 200 folds to 5 × 10 5 CFU/mL bacterial solution.Secondly, prepare the mixture of bacteria and drug with concentration as listed above and incubate the mixture in a shaker at 35℃.At 0, 2, 4, 8, and 24 h, obtain aliquots of 0.1 mL from each tube, serially dilute them in 0.9% sodium chloride, and inoculate 10 µL on LB agar plates.After 24 h of incubation at 37°C, the plates were taken out, and the colonies were calculated (16).Time-kill curves were created by plotting mean colony counts (log 10 CFU/ml) versus time to compare the 24-h killing effects of monotherapy and combination antimicrobial exposure.
The lower limit of detection is 2.0 log 10 CFU/ mL.Bacterial concentrations less than 2.0 log 10 CFU/mL are considered as 2.0 log 10 CFU/mL.If the 24-h concentration is decreased by 2.0 log 10 CFU/mL compared with the most active antibiotic alone, it indicates synergy.Irrelevance is defined as a <2 log 10 CFU/mL increase or decrease at 24 h for the drug combination in comparison with the most active antibiotic alone.Antagonism is defined as a ＞ 2log 10 CFU/mL increase at 24 h for the drug combination when compared with the drug alone (17).

Checkerboard test
The results of the checkerboard combination test are presented in Table 2, and detailed FICI results are listed in Table S2.
Against MBL-producing strains, ATM+CZA displayed remarkable synergy with a rate of 100.00%.Against IMP-producing strains, MEM+CZA a synergy rate of 91.67%, ranking second only to ATM+CZA.Following closely were FOS+CZA and AK+CZA, which displayed synergy rates of 75.00% and 58.30%, respectively.In the case of NDM-pro ducing strains, FOS+CZA and MEM+CZA showed synergy rates of 64.28% and 57.14%, respectively.
In conclusion, the results indicated that MEM+CZA exhibited a strong synergistic effect in all strains, except for NDM-producing strains, where the synergistic rate was only 57.5%.In contrast, the synergistic rate exceeded 90.0% in other strains.Notably, against all MBL-producing strains, which were resistant to CZA and ATM when used individually, the combination of CZA+ATM displayed synergistic effects.Among these combinations, CZA+FOS displayed higher antibacterial activity against MBL-producing strains, with a synergy effect of over 60%, while for KPC producers and non-carbapene mase procedures, the range of synergy rates was 20% to 43%.CZA+AK, CZA+COL and CZA+ERA primarily showed irrelevant effect, with synergy rates ranging from 0.0% to 58.3%.None of the combinations showed antagonism.
The data of SBPI are presented in Table 3. Regarding the combination of CZA+MEM, the median of SBPI is 504, and the mean is 503.75, both of which are higher than other drug combinations among the KPC producers.However, in MBL-producing strains, the SBPI of CZA+MEM ranged from 0.265 to 75, which is lower than the combination of CZA+ATM.This suggests that CZA+ATM may exert a stronger synergistic effect for MBL-producing strains.There is no significant difference in SBPI among the other combinations.

Time-kill assay
The 24 h time-kill results are displayed in Fig. 1.
Against KPC-2-producing isolate L70, antimicrobial combination CZA+AK, CZA+MEM, and CZA+COL, showed a highly synergistic effect.Compared with the most effective single drug showed a decrease of 3 log 10 CFU/ml.Against KPC-2-producing isolate Y070, the combinations of CZA+FOS, CZA+ERA and CZA+MEM showed a stronger synergistic effect, with log 10 CFU/mL decreases of 6.45, 3.29, and 8.47, respectively, compared to the most active component of each single drug.
Against MBL-producing strains Y074 and L013, the combinations of CZA+COL, CZA+MEM, and CZA+ATM both showed synergy effect.For the Y047 strain, which produces IMP enzyme, CZA+ATM was the most effective with log 10 CFU/mL decreases  b "--" indicates that the drug tested in KPC-producing as well as non-carbapenemase-producing strains did not include the AZA.
of 10.43 compared with the control.And for the NDM enzyme producer isolate L13, CZA+MEM showed the greatest effect with log 10 CFU/mL decreases of 10.22.In summary, CZA+MEM showed highly synergistic effect on all seven isolates and ATM+CZA displayed synergy effect in MBL-producing strains.Both of them are consistent with the results of the checkerboard test.

DISCUSSION
The global prevalence of CRKP has gained significant attention in recent years, with numerous studies shedding light on its occurrence across different regions.For example, in a study of Sreeja K (11) CRKP incidence rates in various parts of India ranging from 14% to 69%.Similarly, Yunqing Qiu's study found a prevalence of carbapenem resist ance in pediatric bloodstream infection is 12.79% (18).Another study (19) indicated that about 50% of K. pneumoniae are producers of extended-spectrum beta-lactamase (ESBL).Furthermore, compared to non-ESBL-producing K. pneumoniae, patients infected with these pathogens experience increased mortality, primarily due to a delay in the administration of effective therapy (20,21).Therefore, the rational choice of antimicrobial drugs is crucial.
CZA has been used as a frontline treatment for CRKP infections.However, it is crucial to note the increasing resistance rate to CZA.In a previous study, the overall resistance rate of Enterobacterales to CZA was <0.6%; however the resistance rate increased to 16.7%-21.0%for CRKP (22).Several studies have reported that the resistant rate of CZA against K. pneumoniae was 3.7% (32/872) (23) and 2.5% (24), which are higher than it first been used.Unfortunately, there are few clinical studies on the effects of CZA in combination with other drugs; although some previous studies had reported that about 3.5% cases of CRKP culture positivity persisted when using CZA alone, but the infections were managed with drug combinations (25).Our study revealed that CZA+MEM and CZA+ATM exhibited significant synergistic effects against KPC-2-produc ing and MBL-producing strains, respectively.
Avibactam is a class A and partial class D carbapenemase inhibitor which re-estab lishes susceptibility to carbapenems (26).MEM, a carbapenem antibiotic, binds to penicillin-binding proteins (27), thereby interfering the building of bacterial cell wall and exerting an antibacterial effect.So it's logical that CZA+MEM against KPC-producing isolates showed 91.30% synergy effect (28).However, further researches are warranted to elucidate the underlying mechanisms contributing to the observed 100% synergy effect of CZA+MEM in non-enzyme producing but OmpK35 and OmpK36 porin-deficient isolates.In this study, MEM+CZA also displayed certain synergistic efficiency against MBL-producing strains, especially for IMP-producer, with a synergistic rate of 91.67%.This may be attributed to the fact that the 12 IMP-producing strains showed a lower resistance level to MEM compared to the NDM-producing strains, with 75% of the strains having a MIC of less than 16 µg/mL, but all of the NDM producer had a MIC of 64 µg/mL against MEM.In MBL-producing strains, the synergistic effect between MEM and CZA can be attributed to MEM serving as a substrate for the metalloenzyme, leading to the consumption of the enzyme and a subsequent reduction in its concentration.The provided CZA can inhibit the remaining metal enzymes to achieve a synergistic effect.
ATM is the only beta-lactamases stable to class B metalloenzymes, but it can be hydrolyzed by both class A and C beta -lactamases.In 2014, some researchers (29) had proposed a treatment option of avibactam plus ATM.They hypothesized that avibactam would be able to inhibit the co-existing ESBL and AmpC enzymes; thus, ATM could maintain its antibacterial activity.The study confirmed that CZA+ATM has a good synergy effect for metalloenzyme-producing strains, which is consistent with some previous studies (18).A previous study in China demonstrated that when avibactam is combined, the MIC 50 and MIC 90 of ATM against NDM-producing strains could be reduced by 99.9% and 99.8%, respectively.
Fosfomycin is infrequently used in clinical practice primarily for urinary tract infections since its first discovery (30).Therefore, it has maintained potent antibacterial activity against many multidrug-resistant pathogens including CRE.In addition, the drug has a low protein binding rate and a high tissue penetration rate (31), which give it the potential utility in combination antibiotic drug against MDR pathogens.FOS acts as a time-dependent inhibitor of the MurA enzyme, which catalyzes the first committed step of peptidoglycan synthesis (32).It has a synergy effect with CZA, probably because they act on different times of bacterial cell wall synthesis.
Compared with other studies, we could obtain similar conclusions.Firstly, drug combination is necessary for multi-drug resistant bacterial.For example, to preserve the effectiveness of CZA, its clinical use should be avoided in naturally resistant strains and in those carrying MBLs and certain class D β-carbapenemases (33).Combination therapy plays an important role in these clinical scenarios, as CZA showed a showed synergistic effect when combined with MEM or ATM.Those results are, in part, consistent with those of the previous studies (6,33,34).For KPC-producing strains, combination therapy also holds significant value.Some strains exhibit susceptibility to CZA but may present high MIC values (4 µg/mL or 8 µg/mL), approaching resistance levels.Utilizing CZA alone in treating such strains may lead to treatment failure.It's been well-documented (35) that CZA in monotherapy of critically ill patients infected with carbapenem-producing K. pneumoniae has been associated with higher mortality than in combination therapy.In these scenarios (36), combination therapy can substantially influence the clinical outcome of the patient.
Our study has some advantages: we selected strains with multiple resistance mechanisms including MBL-producing and non-carbapenem-producing strains; in addition to KPC-producing strains, which is the only strains in other studies (37).This study has some limitations such as in vivo experiments are required to confirm our results.Further clinical studies are essential to evaluate the clinical impact of those combinations and establish the efficacy of those regimens in the treatment of infections due to CRKP isolates.

TABLE 1
The characteristics of isolates for time-killing test a