Introduction

In the past decades, due to the heavy use of antibiotics in clinic practice, antibiotic resistance of Acinetobacter baumannii has generally increased worldwide [1, 2]. Of particular concern is the emergence of multidrug resistant (MDR) A. baumannii, the extensive antibiotic resistance nature of which has seriously narrowed the available antibiotic therapy alternatives and made the treatment of bacterial infection caused by MDR A. baumannii a big challenge faced by clinicians [3, 4]. Even worse, nosocomial outbreaks of MDR A. baumannii have been frequently reported in hospitals around the world, which greatly increases the risk of MDR A. baumannii infection and makes the MDR A. baumannii infection a worldwide problem [5, 6]. The wide prevalence of MDR A. baumannii and the high antibiotic resistance feature have made the search for effective antibacterial therapy against MDR A. baumannii an urgent need.

Polymyxins, including colision, polymyxin E and polymyxin B, have currently been regard as the last resort agents for the treatment of MDR A. baumannii [7]. However, the available pharmacodynamic (PD) and pharmacokinetic (PK) data suggest that reliably efficacious plasma concentrations of polymyxin B are difficult to be maintained with polymyxin B monotherapy [7,8,9]. While polymyxins are highly nephrotoxic agents, the likeliness of renal impairment is closely associated with the daily dose of polymyxin [10]. A strategy for promoting antibacterial efficacy without increasing polymyxin exposure is the use of polymyxins in combination with other agents. Actually, this strategy has been recommended by many authorities, such as Clinical and Laboratory Standards Institute (CLSI), and is increasingly used in clinical practice [7, 11, 12]. Theoretically, combination therapy has potential advantages such as increasing the antibacterial efficacy, reducing the dosage of the antibiotics that have high toxicity, and even decreasing the development of antibiotic resistance [13, 14].

Even though a large number of in vitro studies have confirmed synergistic combination of different antibiotics with polymyxin B against different bacterial species, the determination of which antibiotic is more suitable for combination therapy with polymyxin B against MDR A. baumannii remains uncertain. Many studies have examined just one or a few antibiotic combinations, and different laboratories use different analytical methods, as well as different clonal strains, the heterogeneity of the existing data makes it difficult to compare the combination effect of different antibiotics with polymyxin B against MDR A. baumannii [15, 16].

In order to get a systematical evaluation of the synergistic combination of polymyxin B with different antibiotics against MDR A. baumannii, 12 antibiotics that are recommended for routine reporting of antimicrobial susceptibility testing of A. baumannii were selected and examined with checkerboard assays to determine their synergistic rates and synergistic effects with polymyxin B against the MDR A. baumannii isolated from a Chinese tertiary hospital.

Materials and methods

Bacterial isolates and chemicals

A total of 128 non-duplicate MDR A. baumannii isolates were obtained from various specimens of inpatients in Shiyan Renmin Hospital, a Chinese tertiary hospital that located in the center area of China, during the period from 2018 to 2020. Species that identified with the Vitek2 compact system (BioMerieux, France) were further confirmed with MALDI-TOF mass spectrometry (Bruker, China). Antibiotic susceptibility of those isolates referred to the clinic antimicrobial susceptibility reports and MDR was defined as being resistant to at least three of the six antibiotic classes, including β-lactam/β-lactamase inhibitor combinations, cephalosporins, carbapenems, aminoglycosides, tetracyclines, fluoroquinolones. A. baumannii ATCC19606 was used as the control strain. All antibiotics employed in this study were purchased from Solarbio (Beijing, China) and dissolved with sterile water to prepare stock solution.

Multilocus sequence typing and clonal complex analysis

32 out of 128 isolates of MDR A. baumannii were randomly selected and cultivated with LB broth for 24 h. Then bacteria were collected and total DNA was extracted with Bacterial Genomic DNA Extraction Kit (SAINT-BIO, Shanghai). Multilocus sequence typing (MLST) of MDR A.baumannii was determined by amplifying seven housekeeping genes as described previously, including cpn60, fusA, gltA, pyrG, recA, rplB, rpoB. The fragments containing seven housekeeping genes were sent to Biomed (Beijing, China) for purification and DNA sequencing service. The numbers of alleles and sequence types (STs) were assigned by comparing with MLST online database (https://pubmlst.org). The goeBURST algorithm was used to infer the evolutionary relatedness of the STs.

Susceptibility testing

MICs of polymyxin B and 12 other antibiotics were determined with the broth microdilution method according to the CLSI guideline (https://clsi.org/). Briefly, bacterial cell suspensions prepared with cation-regulated Mueller–Hinton broth (CAMHB) were dispensed into the wells of 96-well microtiter plate and made the final bacterial concentration of 5 × 105 CFU ml−1 for each well. Polymyxin B and 12 other antibiotics were separately added into each well to get desired concentration, ranging from 256 to 2 μg ml−1 with twofold dilution for piperacillin-tazobactam (TZP) and sulfamethoxazole (SXT), 128 to 1 μg ml−1 with twofold dilution for cefperazone-sulbactam (SCF), ceftazidime (CAZ) and amikacin (AK), 64 to 0.5 μg ml−1 with twofold dilution for cefepime (FEP), 32 to 0.25 μg ml−1 with twofold dilution for tobramycin (TOB), minocycline (MH), imipenem (IPM) and meropenem (MEM), 16 to 0.125 μg ml−1 with twofold dilution for levofloxacin (LEV) and polymyxin B (PB), and 8 to 0.0625 μg ml−1 with two-fold dilution for ciprofloxacin (CIP). The plates were read after incubating for 24 h, and MIC was defined as the lowest concentration of antibiotic required to visibly inhibit the growth of bacteria.

Checkerboard studies

Basing on the susceptibility testing results, checkerboard assays were further conducted to determine the pharmacodynamic interaction of polymyxin B with 12 other antibiotics on antibacterial activity. In this experiment, polymyxin B and each other antibiotic were prepared by two-fold serial dilutions and mixed to produce different concentration combinations. The results were read after incubated for 24 h and interpreted as follows: FICI ≤ 0.5, synergistic; 0.5 < FICI ≤ 4.0, indifferent; and FICI > 4.0, antagonistic.

Synergistic combination analysis

Synergistic combination was analyzed from two aspects, one was synergistic rate and the other was synergistic effect. For each antibiotic combination, synergistic rate was calculated as the percentage of the isolates with FICI ≤ 0.5.

Results

Molecular epidemiological characteristics of MDR A. baumannii

Two STs have been identified with 32 isolates of MDR A. baumannii selected for MLST analysis. ST92 was the dominant ST, taking up to 71.9%, and its single locus variant (SLV) ST75 accounted for the remaining 28.1%. Both of those two STs were grouped into the clonal complex (CC) CC92, one of the largest CC for A. baumannii in the database of PubMLST.

Antibiotic resistance profiles of MDR A. baumannii

Antibiotic resistance rates of 128 MDR A. baumannii isolates to PB and 12 other antibiotics are listed in Table 1. The data showed that MDR A.baumannii obtained from this hospital exhibited high resistance rates to TZP, SCF, CAZ, FEP, IPM, MEM, TOB, CIP, LEV and SXT (100% to all). The percentage of the isolates that were resistant to AK and MH were 65.6% and 29.7% respectively, and no isolate appeared to be resistant to polymyxin B. Comparing the antibiotic resistance profile of ST92 and ST75, great similarity could be found, with both of those two STs showing high antibiotic resistance rates to TZP, CAZ, SCF, FEP, IPM, MEM, CIP, LEV, TOB and SXT (100% to all), and low to MH (26.1% for ST92, 33.3% for ST75) and PB (0% for both). Notably, significant difference in antibiotic resistance rate to AK could be detected between those two STs, with ST75 showing much higher rate than ST92, with 88.9% vs 47.8%.

Table 1 Synergistic combinations of polymyxin B with 12 other antibiotics against MDR A. baumannii obtained from a Chinese tertiary hospital

Synergistic combinations of polymyxin B with 12 other antibiotics against MDR A. baumannii isolated in one hospital

Synergistic combinations of polymyxin B with 12 other antibiotics against MDR A. baumannii isolated in one hospital were analyzed with the checkerboard assay. The results showed that SCF had the highest synergistic rate with polymyxin B, reaching up to approximate 75.8%, followed by MH, IPM and MEM, with about 69.5%, 45.3% and 39.1% respectively. CAZ had the synergistic rate of 31.2% with polymyxin B, while FEP, AK and SXT had just 24.2%, 17.2% and 4.7%, respectively (as shown in Table 1). No synergism was noted in combinations of polymyxin B with TZP, CIP, LEV and TOB, with any isolate.

Among the seven antibiotic combinations that have a synergistic rate of >10%, SCF showed the best synergistic effect with polymyxin B, with the antibiotic susceptibility raised by approximate 10.9-fold by 1/4 MIC of polymyxin B. AK, IPM and MEM had similar synergistic effect, with the antibiotic susceptibility raised by 9.5-, 9.6- and 8.4-fold by 1/4 MIC of polymyxin B, respectively (as shown in Table 1). Though CAZ had lower synergistic rate with polymyxin B than MH, it had much better synergistic effect with polymyxin B, with the antibiotic susceptibility raised by approximate 8.3-fold by 1/4 MIC of polymyxin B, compared with 6.3-fold for MH. FEP had relative poorer synergistic effect with polymyxin B among the seven combinations, with the antibiotic susceptibility raised by just 4.8-fold by 1/4 MIC of polymyxin B.

The checkerboard assay has also revealed that none of these 12 antibiotics had antagonistic effect with polymyxin B against MDR A.baumannii isolated in this hospital.

Discussion

Like many healthcare facilities around the world, many hospitals in China have also experienced worrysome nosocomial infections, particularly that caused by MDR A. baumannii [17, 18].

In this study, MLST analysis has revealed that the genetic background of MDR A. baumannii obtained from a Chinese tertiary hospital is quite simple, with just two sequence types (92ST and 75ST) that show close evolutionary relationship and belong to the same clonal complex (CC92). Also, antibiotic susceptibility analysis has shown that those two STs have great similarity in their antibiotic resistance profiles, both of them with high resistance rates to antibiotics including TZP, CAZ, SCF, FEP, IPM, MEM, LEV, CIP, TOB, and SXT, and low-resistance rates to MH and PB. Both the molecular epidemiological characteristic and antibiotic susceptibility related phenotype suggest that nosocomial spreading and nosocomial infection of MDR A. baumannii have occurred in this hospital.

With the aim to explore the effective antibacterial strategy against MDR A. baumannii as characterized in this hospital, synergistic combination of polymyxin B with 12 other antibiotics were analyzed with MDR A. baumannii obtained from this hospital. The results have demonstrated that SCF has the highest synergistic rate with polymyxin B, reaching up to 75.8%, followed by MH, IPM, MEM, CAZ, FEP, AK and SXT. Also, SCF has the best synergistic effect with polymyxin B, with the susceptibility of SCF that in synergistic combination with polymyxin B increased by about 10.9-fold by 1/4 MIC of polymyxin B, that is followed by IPM, AK, MEM, CAZ, MH and FEP.

Synergistic combination of polymyxin B with some antibiotics as examined in this research has also been reported in various studies with A. baumannii strains obtained from different regions of the world. Some are in agreement with our study. For example, synergistic combination of polymyxin B with IPM has been confirmed in two studies with A. baumannii obtained from New York, UK and Isparta, Turkey, and with MEM in two studies with A. baumannii obtained from Isparta, Turkey and São Paulo, Brazil, respectively [19,20,21]. Also, synergistic effect of polymyxin B with MH has been demonstrated in a study with A. baumannii obtained from Jinan, China, and with CAZ and AK in a study with A. baumannii obtained from São Paulo, Brazil, respectively [22, 23]. Giving that A. baumannii strains obtained from different geographical regions are likely to have different molecular epidemic characteristics, synergistic combinations confirmed with A. baumannii strains obtained from different cities can be assumed to be specific and charcateristic to the area [24,25,26,27].

However, some reports are different from our study. For example, even though high synergistic rate of polymyxin B with SCF against MDR A. baumannii has been demonstrated in this study, synergistic effect between those two antibiotics could not be detected with any tested isolate in a study with MDR A. baumannii obtained from Kars, Turkey [28]. Also, though the synergistic combination of polymyxin B with CIP failed to be confirmed in this study, a synergistic rate of 14.29% for this combination had been obtained in a previous study with 21 MDR A. baumannii isolates obtained from Annaba, Algeria [29]. The discrepancy between those studies may due to the different molecular characters of A. baumannii strains obtained from different regions, and suggests that synergistic combinations of those antibiotics with polymyxin B may be bacterial strain related. For a better understanding of synergistic combination of polymyxin B with those antibiotics against MDR A. baumannii, further studies with well characterized MDR A. baumannii strains obtained from different regions may be required.

In a previous study, meta-analysis was performed, by incorporating all published data about in vitro synergistic combinations of polymyxin B with different antibiotics against A. baumannii obtained from different regions of the world, to evaluate the general synergistic rate of polymyxin B with those antibiotics against A. baumannii [30]. Consistent with our results, the general synergistic rate of carbapenems with polymyxin B in this meta-analysis was approximate 39.6%, compared with 45.3% for IPM and 39.1% for MEM in our study. Two possible reasons may account for this similarity, one is that the combination of polymyxin B with carbapenems has general synergistic effect against the A. baumannii strains occuring all over the world, with the synergistic rate of approximate 39.6%, the other is that A. baumannii strains used in this study may represent the most epidemic clonal complex around the world. Actually, the clonal complex that the A. baumannii strains in this study belong to is the main clonal complex for A. baumannii in Pubmlst dataset, and also the epidemic clonal complex that is widely spreading in Asia, Europe and America [31,32,33,34].

Basing on the result of this meta-analysis, carbapenems were proposed as the best partner for combination therapy of polymyxin B against A. baumannii, as they had relatively higher synergistic rates with polymyxin B compared with the other antibiotics included in analysis [30]. However, in this study, we have found that polymyxin B has the higher synergistic rate and the better synergistic effect with SCF than with IPM, with 75.8% vs 45.3% in synergistic rate and 10.9- vs 9.6-fold increased in susceptibility by 1/4 MIC of polymyxin B, than with MEM, with 75.8% vs 39.1% in synergistic rate and 10.9- vs 8.4-fold increased in susceptibility by 1/4 MIC of polymyxin B, respectively. Though MH has a relatively poorer synergistic effect with polymyxin B compared with carbapenems, with just about 6.3 fold increased in susceptibility by 1/4 MIC of polymyxin B, the synergistic rate of MH with polymyxin B is much higher than that of IPM, with 69.5% vs 45.3%, and than that of MEM, with 69.5% vs 39.1%, respectively. Those results suggest that SCF and MH may be more suitable for polymyxin B based combination therapy against MDR A. baumannii than carbapenems and further researches are deserved to determine their clinical value on the treatment of MDR A. baumannii infection.

Cephalosporins, including CAZ and FEP, were also included into our study and showed synergistic rates of 31.2% and 24.2% with polymyxin B against MDR A. baumannii respectively. Though SXT had a relatively lower synergistic rate of 4.7%, the synergistic effect of it with polymyxin B against MDR A. baumannii was first determined in this study. All of the data have potential value on the treatment of MDR A. baumannii infection. Except with AK, where it has been shown to have a synergistic rate of 17.2% with polymyxin B against MDR A. baumannii, giving that both AK and polymyxin B have nephrotoxicity and the combination of these two antibiotics are likely to enhance the side effect, this combination is not recommended to be used in clinical practice.

In summary, our study gives a complete analysis of the synergistic combinations of polymyxin B with 12 different antibiotics against MDR A. baumannii, which could provide some useful insights for the selection of antibiotics that are most suitable for combination therapy of polymyxin B. Also, some synergistic combinations against MDR A. baumannii, to the best of our knowledge, are first noted in our study, which could provide additional therapeutic alternatives for the combination therapy of polymyxin B against MDR A. baumannii.

There are two limitations of this study that should be considered. One is that this study was just based on the isolates of MDR A. baumannii obtained from one hospital of China. It seems that these isolates have a close evolutionary relationship, similar antibiotic resistance profiles and related antibiotic-resistant mechanisms. Strains obtained from different regions may have different molecular epidemic characteristics and different conclusions may be obtained with strains obtained from different regions. For this concern, further studies with different MDR A. baumannii strains are required. The other is that this study has just evaluated the in vitro synergistic combination of polymyxin B with 12 other antibiotics. However, the question that which combination of antibiotic concentration should be employed for clinical use of those antibiotic combinations is still uncertain. Giving that different combinations of antibiotic concentration influence not only the antibacterial efficacy but also the side effect on patients, so it is necessary to make it clear which combination will benefit the patients the best. Furthermore, for each combination therapy, not only the antibacterial activity, but also the toxicity and adverse effects should be considered. One of the most serious adverse effects of polymyxin B therapy is the does related nephrotoxicity, so any antibiotic that can aggravate the nephrotoxicity of polymyxin B should not be used in combination with it. For optimizing those polymyxin B based combination therapies, further pharmacokinetic and pharmacodynamic studies are required to determine the appropriate combination of antibiotic concentration and to analyze the toxicity profile of each combination. Also, animal models and clinical trails are required to evaluate the clinical value of those combinations in the treatment of MDR A. baumannii infection.