Introduction
Klebsiella pneumoniae is an increasingly important bacterial pathogen that causes severe lift-threatening diseases [
1]. However, data from China Antimicrobial Surveillance Network (CHINET) indicated the resistance rate to imipenem in
K. pneumoniae isolates has increased from 0.4% in 2005 to 25.0% in 2018 [
2,
3]. The increasing emergence of carbapenem-resistant
K. pneumoniae (CRKP), especially
Klebsiella pneumoniae carbapenemase (KPC)-producing
K. pneumoniae (KPC-Kp), has become an urgent public health problem in healthcare settings, resulting in higher morbidity, mortality and medical cost [
4,
5]. The risk of high mortality related to these infections was inappropriate empirical antimicrobial treatment [
6]. However, the paucity of new classes of antibiotics with which to treat such circumstance has led to regain significant interest in the revival of fosfomycin (FOS), tigecycline (TGC), and colistin methanesulfonate (CMS) as last-resort drugs [
7]. Therefore, evaluation of the efficacy of these alternative options is necessary to manage the immediate threat of CRKP in the the ‘bad bugs, no drugs’ era, in addition to facilitating the development and clinical authorization of novel antimicrobials [
8]. As an in vitro susceptibility is insufficient to choose rational antibiotic or dosing regimens in clinic, the introduction of population pharmacokinetics (PK) with monte carlo simulation (MCS) integrates population-PK parameters and population-minimum inhibitory concentration (MIC) pathogen data together to calculate the likelihood of achieving a certain target [
9]. This approach may be applied to optimize dosing regimens, maximize the desired effects, and re-evaluate reasonable clinical breakpoints.
Colistin (CST) and TGC show favourable in vitro activity against CRKP [
10]. However, the role of CST and TGC in the treatment of severe nosocomial infections remains controversial. The reason may be a great inter-individual variability in the population PK and heteroresistance for CMS. And for TGC, a large volume of distribution and low concentrations in blood, urine, and epithelial lining fluid of the lungs were observed [
11‐
13]. Some experts have revealed the current recommended dosage of TGC and CMS may be suboptimal, and higher doses should be considered [
14,
15]. Furthermore, several studies have shown that combination therapy resulted to the promising outcome than monotherapy in combating multidrug-resistant infections, and the dosing regimens included TGC or CMS were associated with lower mortality [
16]. And for FOS, it may remain active against a considerable proportion of CRKP, especially for carbapenem-resistant
Enterobacteriaceae [
10]. It can be used in the management of difficult-to-treat infections combined with other antimicrobial agents [
17]. There is confusion regard to whether FOS displays time- or concentration-dependent bactericidal activity [
18,
19]. It seems that this depends on the microorganism under study. Therefore, two different estimations of PK/PD indices for FOS may be done in our analysis. Although several studies regarding the MCS of FOS, TGC and CMS have been done, most of these evaluations were evaluated primarily in a monotherapy setting, and combination antimicrobial synergy studies using this method are scarce.
To date, there have been limited studies concerning on the optimal dosage regimens of the three antibiotics for the treatment of KPC-Kp infections in our region. The aim of our study was to: (i) re-evaluate reasonable clinical breakpoints of FOS, TGC and CST using MCS; (ii) assess the efficacy of three candidate antibiotics against KPC-Kp by mono- or combination therapy; (iii) find the prompt initiation of appropriate antimicrobial therapy against KPC-Kp.
Discussion
KPC-Kp are increasingly prevalent and has been becoming a global public health concern. This dilemma often resulted in early inappropriate antimicrobial therapy associated with a high-risk factor for the mortality rates [
6]. The KPC-Kp, used in our study, were highly susceptible to CST, nevertheless showed limited susceptible to TGC and FOS. The reason may be the widely use of TGC in the treatment of Carbapenem-resistant Gram-negative bacterial infections in China. In this regard, it is critical to know local trends in resistance and population-MIC distributions in order to achieve better empirically therapeutic outcomes [
39]. The adequate empirical antibiotic treatment should be considered local and recent data on antimicrobial resistance as well as inter-individual variation of PK behavior in virtual patients. Bayesian-based dosing for patients was conducted in our study to provide individualised dosing regimens from a patient’s own PK parameter estimates. Thus, a truly optimized regimen could be derived for each patient. This can improve the clinical cure rate, especially in critically ill patients infected with highly resistant pathogens, rather than the manufacturer’s prescribing information. To the best of our knowledge, this is the first and largest study to estimate the combined treatment of FOS with TGC or CMS against KPC-Kp using population-PK model in China. Our findings highlighted the importance of high dose TGC or CMS in combination with FOS against KPC-Kp. This would be useful in empirically treating patients infected with KPC-Kp or high risk factors of CRKP, as quite a few tertiary and secondary health care settings failed to afford the MIC results of FOS and CST in clinic.
The patients with CrCL of < 30 ml/min was not simulated in our study as such patients in the stage of end stage renal disease (ESRD) often require dialysis therapies. Several changes of antibiotics in absorption, distribution and metabolism would be noted after dialysis [
40]. This depends on the characteristics of dialyzing membrane and drug, the rate of blood flow as well as the duration of therapy [
40]. The simulated CrCL of > 30 ml/min was in accordance with the reported CrCL in critically ill patients [
41].
FOS is being used frequently against multidrug-resistant organisms. Our data revealed that none of the FOS regimens in monotherapy was able to achieve PK/PD targets related to antimicrobial efficacy for KPC-Kp. Consistent with another PK study of FOS 8 g q8h in critically ill patients, a mean of C
max 307 mg/L also failed to reach the target because of the high MICs [
42]. Fortunately, the combination with TGC brought the FOS MIC
90 to ≤64 mg/L, and thus, providing sufficient antimicrobial coverage against KPC-Kp. The CFRs of combination therapy were raised to > 80% in normal renal function and > 90% in renal impairment based on the PK/PD target of AUC
24/MIC. Thus, empirical therapy in the treatment of infections caused by KPC-Kp with high MICs can use the combination regimens of FOS and TGC. Besides, drugs in combination could completely suppress all clones resistant to FOS at a low dose of 12 g/day [
43]. Although the FOS daily dose of 18 g to 24 g in combination with TGC might be promising, these high doses may cause adverse side effects, such as hypokalemia and saline overload [
44]. It is worth noting that it is still not fully elucidated if dose adjustment is needed for the CrCL of 40 to 80 ml/min. For patients with CrCL < 40 ml/min, a reduction of daily recommended dose is proposed [
45]. As the means of simulated CrCL in our study was 40–55 ml/min for the decreased renal function cohort, these high doses in combination may be a safe and effective therapeutic method for management of difficult-to-treat infections in such patients.
TGC showed limited in vitro activity against KPC-Kp. The data of TGC MIC, used in the MCS, was relatively high compared with other studies [
10,
46]. Thus, the recommended standard dosing regimen of TGC (100 mg loading dose followed by 50 mg q12h) failed to achieve PK/PD targets for KPC-Kp in combination therapy. Currently, the role of TGC in treating critically ill patients is still controversial [
12]. In 2013, the Food and Drug Administration (FDA) reported an increased risk of death associated with TGC use [
47]. The reason may be the suboptimal dosing regimens and relatively high MICs in certain bacterial strains [
48,
49]. Yamashita. et al. indicated peak TGC serum levels were low (0.63–1.4 mg/L) after administrating the standard dosing regimen of TGC [
50]. Thus, it is still essential to evaluate the efficacy of TGC dosing regimens owing to the above situation and limited treatment options. Consistent with previous studies, our findings indicated that standard TGC dosing regimen was suboptimal, while an increase of the daily dose could achieve better PTA and CFR [
46]. High dose has been evaluated in the treatment of CRKP and found lower mortality and better clinical responses compared with the recommended standard dosage [
51,
52]. Due to the long t
1/2 (42 h following multiple doses) and linear PK characteristics of TGC, once daily high dose TGC regimens were also simulated in our study and reached favourable CFR in combination therapy. Thus, its clinical value as an option of last resort for treating multidrug-resistant isolates is worthy of exploration. In view of these results, high dose is essential to obtain maximum concentration-dependent killing, especially for Carbapenem-resistant organisms with an MIC of 2 mg/L. But the incidence of adverse events, mainly concerning gastrointestinal disorder, was elevated in the high TGC group [
51]. Of note, the difference in serious adverse events was not statistically significant. TGC is well tolerated at high dose. Similar clinical outcomes of high-dose vs low-dose TGC were also described in a meta-analysis study, including 1041 patients [
53]. It has been suggested that no dose adjustment was required for TGC in renal or hepatic impairment, unless there is severe hepatic dysfunction. From our simulated results, TGC 200 mg loading dose followed by 100 mg q12h in combination with FOS 8 g q8h in normal renal function or FOS 4 g q8h in renal impairment might be reasonable in empirically treating critically ill patients infected with KPC-Kp, so as to maximize a favorable clinical response and minimize exposure-related toxicity. In the future, well-designed studies especially randomized controlled trials (RCTs) are required to establish the effectiveness and safety of high-dose TGC.
The clinical breakpoint of CST at present is 2 mg/L for
Enterobacteriaceae. However, in such situation, only the two loading regimens achieved PTA values higher than 90% in the renal impairment. This would be expected to increase the likelihood of acute renal failure. Presently, the daily dose is suggested to be reduced in patients with decreased renal function. Our findings showed that CMS dosing regimens in combination with FOS led to a CFR in the range of 60–80% and 80–92% for normal renal function and renal impairment, respectively. Similar clinical cure rate with no significant renal toxicity was observed in patients with sepsis due to Gram-negative bacteria susceptible only to CST and treated with 4.5 million IU q12h [
54]. However, lower clinical cure rates (57–75%) have been reported in the low CMS dose (2.2–6 million IU/day) group [
55,
56]. Such low daily doses always failed to produce sufficient drug exposures to reach the PK/PD target for isolates with an MIC of 0.5–1 mg/L in our study. It has been stated that the current recommended dose of CMS by manufacturers is associated with suboptimal concentrations in a large number of the patients [
57]. Worsely, such sub therapeutic concentrations often resulted to the amplification of colistin-resistant subpopulations in heteroresistant strains [
58]. Combination therapy is still needed in view of our findings and previous studies. Moreover, a previous meta-analysis indicated that mortality was significantly higher with polymyxin monotherapy compared with combination therapy with TGC, FOS or aminoglycosides, especially for
K. pneumoniae bloodstream infection [
59]. Considering that increasing use of CMS and the spread of
mcr-1 gene in plasmid might be leading to the emergence of CST resistance worldwide [
60,
61], the combination with FOS can take into account the antibacterial efficacy and the reduced CMS daily dose so as to decrease the likelihood of the risk of nephrotoxicity, which is instructive in managing patients with decreased renal function.
There are several limitations to this study. First, the population PK model of FOS was developed form 12 enrolled patients with a total of 515 plasma samples [
27]. And for TGC and CMS, 146 and 105 patients were included in their studies [
29,
31]. Thus, the rich PK properties of FOS was not fully evaluated, meaning that other relevant covariates might not be included in the model. Second, the MICs of the KPC-Kp populations isolated from the three hospitals may not be representative of the MIC distributions in other regions. Third, a precise prediction of the efficacy of antibiotics against KPC-Kp is challenging because of the complicated condition in critically ill patients. Although the host immune response was not evaluated in our study, the presence of a competent immune system can markedly increase the efficacy of drugs against bacterial infections [
62]. Moreover, combination therapy was often used for such patients in clinical practice. In addition, the PK/PD targets used in our study might be not fully elucidated, as the PK/PD targets used in our study were established for monotherapy. Further studies, including PK/PD simulations, animal models, and clinical trials, are urgently needed to evaluate the efficacy and toxicity of FOS, TGC and CMS against CRKP.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.