Role of Double-Carbapenem Regimen in the Treatment of Infections due to Carbapenemase Producing Carbapenem-Resistant Enterobacteriaceae: A Single-Center, Observational Study.

Purpose
(i) To compare infections caused by carbapenem-susceptible (CS) and carbapenemase producing carbapenem-resistant Enterobacteriaceae (CP-CRE); (ii) to evaluate the clinical effectiveness of the double-carbapenem (DC) regimen in comparison with the best available treatment (BAT) in infections caused by CP-CRE; and (iii) to determine the exact minimal inhibitory concentrations (MICs) of meropenem/ertapenem (MEM/ETP) and the degree of in vitro ETP+MEM synergism in subjects receiving the DC.


Methodology
Over a 3-year period (2014-2017), patients with infections due to Enterobacteriaceae were included in a single-center, retrospective, observational study. According to the susceptibility to carbapenems, subjects were divided into CSE and CP-CRE groups. CP-CRE group was further divided into subjects receiving the DC regimen and those treated with other regimens (BAT group). Clinical characteristics and the presence of 5th-day response and 60-day outcome were evaluated for DC and BAT groups. The determination of MEM and ETP actual MICs and the MEM+ETP synergistic activity were performed on strains obtained from subjects receiving the DC regimen.


Results
A total of 128 patients were included in the study: 55/128 (43%) with infections due to CP-CRE and 73/128 (57%) with infections due to CSE. Among CP-CRE (n=55), 21 subjects (39%) were treated with the DC regimen whereas 34 (61%) received BAT. No differences in terms of severity of infection, presence/absence of concomitant bacteremia, type of infection, and resolution of infection were found; in contrast, DC group tended to have a higher rate of sepsis or septic shock at the onset of infection and a higher rate of 5th-day response. MICs 50/90 were 256/512 and 256/256 μg/mL for MEM and ETP, respectively. Overall, complete in vitro synergism was found in 6/20 strains (30%).


Conclusion
The DC regimen is a valid and effective therapeutic option in patients with infections due to KPC producing CRE, including those with bacteremic infection and more severe clinical conditions. The clinical effectiveness is maintained even in the presence of extremely high MEM MICs.


Background
The rapid spread of multidrug-resistant bacteria has become a public health concern, especially in some countries where the spread of carbapenem-resistant microorganisms is endemic [1]. In particular, infections caused by CP-CRE are associated with a high treatment failure and consequent high mortality, given the limited therapeutic options and the lack of worldwide availability of new drugs such as ceftazidime/avibactam [2].
Risk factors for CP-CRE infections have been widely investigated and serve as possible drivers of prompting an appropriate antimicrobial therapy, aiming at improving the infection cure and reducing mortality [3].
Although the combination therapy is preferred over monotherapy, the optimal management of CP-CRE systemic infections remains a real challenge, which seems even more complicated given the emergence of resistance to ceftazidime/avibactam [4] and the rising diffusion of strains harbouring enzymes other than carbapenemases [1]. Recently, several efforts have been made with the aim of finding the most appropriate antimicrobial regimen according to the susceptibility profile of the microorganisms and the severity of infection [5,6]. In this setting, the double-carbapenem regimen retains a place in therapy in patients with high risk of mortality, pandrug resistant organisms, and lack of therapeutic options [5].
Based on these premises, aims of the study were (i) to compare infections caused by CSE with those caused by CP-CRE, (ii) to evaluate the clinical effectiveness of the DC regimen in comparison with BAT in infections caused by CP-CRE, and (iii) to determine the exact MICs of MEM/ETP and the degree of in vitro ETP+MEM synergism in subjects receiving the DC.

Study Population.
This was a single-center, retrospective, observational study including patients hospitalized over a 3year period (2014-2017), at the Department of Public Health and Infectious Diseases (Sapienza University, Rome) with infections due to Enterobacteriaceae.
Carbapenem susceptibility was determined using VITEK-2 system and interpreted in accordance with EUCAST breakpoint [6] whereas CP-CRE were defined following CDC case definition [7].
Accordingly, subjects were divided into 2 groups: CP-CRE and CSE. CP-CRE group was further divided into subjects receiving the DC regimen and those treated with other regimens, defined as BAT group (Figure 1).
The DC consisted of ertapenem (1 g/day) followed by high doses of meropenem (6 g/day) or modified according to creatinine clearance. BAT was defined as the definitive therapy chosen by the Infectious Diseases specialists according to susceptibility profile of the microorganisms and the clinical conditions of the patients.
Demographic, clinical, and laboratoristic parameters were collected for each subject. Inclusion criteria were age >18 and patients with infections due to Enterobacteriaceae receiving antimicrobial therapy. Isolates collected from all sites of infection were also included. However, in case of multiple cultures from the same patient, only the first isolate causing infection was considered in the study. Exclusion criteria were age <18 and pregnancy.
Given the unconventionality of the treatment, all study participants receiving the DC regimen gave informed written consent. The study was approved by the local Ethics Committee.

Definitions.
The clinical presentation of infection (sepsis, sepsis shock) was defined in accordance with the international guidelines [8].
The clinical and/or microbiological response at day 5 was defined as resolution of signs and symptoms of infections (i.e., defervescence, improvement of clinical conditions and imaging upon antimicrobial treatment) and/or negativity of cultures performed after 5 days of antimicrobial treatment, respectively, and expressed as a nominal variable. In addition, time to clinical response was defined as time (days) to resolution of fever and improvement in clinical or radiological status, expressed as a continuous variable. As for outcomes, clinical cure was defined as survival at 60 days, resolution of signs and symptoms of infection, and absence of recurrence at 60 days following the onset of infection [9]. Infection relapse was defined as recrudescence of infection after an initial response [10].

Microbiological Studies.
The antimicrobial susceptibility pattern of Enterobacteriaceae was obtained through the VITEK-2 system (bioMerieux, Marcy l'Etoile, France).
Strains obtained from subjects receiving the DC regimen underwent additional microbiological analyses, including the phenotypic determination of carbapenemases [11]. The determination of meropenem and ertapenem actual MICs was obtained by the macrobroth dilution method [12] whereas the synergistic activity of meropenem plus ertapenem was performed by the checkerboard method and the fractional inhibitory concentration index (FICI) calculation. Briefly, a 96-well microtitre plate containing antibiotic combinations at different concentrations and a final inoculum of ∼5 x 10 5 CFU/ml of CP-CRE was incubated at 37 ∘ C for 24h under static conditions in Mueller Hinton Broth. The FICI of each combination was defined as follows: ∑FIC: FICA + FICB= MICA+B/MICA alone + MICB+A/MICB alone. A FICI ≤0.5 indicated synergism [13]. Experiments were performed in triplicate and the results were averaged.

Statistical Analysis.
Results were expressed as mean ± standard deviation (SD) or median (range) and as percentages for continuous and categorical variables, respectively. Categorical variables (such as clinical and/or microbiological response at day 5) were compared by using the X 2 or Fisher's exact tests, as appropriate, whereas continuous data (such as time to clinical response) were analyzed with Student's ttest and the nonparametric Mann-Whitney test. Statistical analyses were performed using STATA 9 software (STATA Corp. LP, College Station, Texas, USA) and GraphPad Prism version 7 for Windows (Graphpad Software MacKiev), as appropriate. All statistic tests were 2-tailed and a p value <0.05 was considered statistically significant.

Carbapenemase Producing Carbapenem-Resistant vs
Carbapenem-Susceptible Infections. A total of 128 patients were included in the study: 55/128 (43%) with infections due to CP-CRE and 73/128 (57%) with infections due to CSE (Table 1). Although not statistically significant, length of hospitalization before the onset of infection was higher in CP-CRE than in CSE (median 29.5 vs 17 days, p=0.13).
Clinical presentation was more severe (sepsis and/or septic shock) in CP-CRE than in CSE [

Microbiological Analyses.
Microbiological analyses were performed on strains collected from subjects receiving DC regimen (n=20; 1 strain was not available) and are represented in Table 3.
All the isolated CP-CRE harboured KPC enzymes, which is in accordance with the local epidemiology [14].
Among subjects with meropenem MIC was ≤128 g/mL (n=7), which has been found as the best in vitro MIC value for predicting the highest activity of the DC [15] the clinical outcome at 60 days was cure or relapse in the totality of cases (5 cure, 2 relapse) whereas in patients with meropenem MIC >128 g/mL (n=13) death occurred in 2 cases and cure in 11 (Table 4). Table 1: General characteristics of study population, according to the carbapenem susceptibility of Enterobacteriaceae. CP-CRE: carbapenemase producing carbapenem-resistant Enterobacteriaceae; CSE: carbapenem-susceptible Enterobacteriaceae;COPD: chronic obstructive pulmonary disease; HA: hospital-acquired; CA: community-acquired. ∘: Only subjects with active urinary tract infection requiring antimicrobial therapy were included; §: defined as resolution of signs and symptoms of infections (i.e., defervescence, improvement of clinical conditions and imaging upon antimicrobial treatment) and/or negativity of cultures performed after 5 days of antimicrobial treatment, respectively. * * : sepsis and septic shock were defined according to international guidelines [8].  Table 2: Comparison between subjects treated with the double-carbapenem regimen (DC) and those treated with the best available treatment (BAT). * : two infections were present in some patients. ∘: only subjects with active urinary tract infections requiring antimicrobial therapy were included; §: defined as resolution of signs and symptoms of infections (i.e., defervescence, improvement of clinical conditions and imaging upon antimicrobial treatment) and/or negativity of cultures performed after 5 days of antimicrobial treatment, respectively. * * : sepsis and septic shock were defined according to international guidelines [8].

Discussion
Infections caused by CP-CRE are characterized by a higher morbidity and mortality than those caused by carbapenemsensitive strains [2]. Given the worldwide spread of CP-CRE and the growing emergence of resistance to antimicrobials such as colistin and aminoglycosides, which have been used as last resort drugs, there is a growing literature investigating the best therapeutical regimen according to prognostic scores and/or antimicrobial susceptibility pattern of the microorganisms [16].
Furthermore, new agents with activity against CP-CRE show preferential activity against certain type of carbapenemases [17] and unfortunately their availability is still restricted to some countries, with obvious therapeutic limitations. The recent use of drugs such as ceftazidime/avibactam led to the consideration that it might be considered as a valid option in the setting of CRE infection [17]; however, its use might be undermined by the emergence of resistance, especially in strains harbouring KPC-3 enzymes and even during treatment [18]. Table 3: Microbiological studies on strains isolated from patients treated with the double-carbapenem regimen and correlation with clinical outcome. MEM: meropenem; ETP: ertapenem. ∘: complete synergy was defined as FICI ≤ 0.5, indifference as FICI > 0.5-4.0, and antagonism as FICI > 4.0 [13]. * : one strain was not available for additional microbiological studies. NA: not applicable.
In the present study, all consecutive patients with infections caused by Enterobacteriaceae hospitalized at the Department of Public Health and Infectious Diseases (Sapienza University, Rome) over a 3-year period were included. Apart from the observed differences between CP-CRE and CSE infections, which confirmed the widely reported data in the literature regarding epidemiology (with K. pneumoniae being the most frequent CR-CPE), a more severe clinical presentation and a lower rate of infection cure in CP-CRE, we were able to analyze a consistent number of patients treated with the DC regimen in comparison with the BAT group. As a matter of fact, no differences in terms of severity of infection, presence/absence of concomitant bacteremia, type of infection, and resolution of infection were found; in contrast, subjects treated with the DC tended to have a higher rate of sepsis/septic shock at the onset of infection and a higher rate of 5 th -day response. Taken together, these findings confirm that the DC regimen represents a valid therapeutic option when no other alternatives are possible, with a global high clinical cure, similar to that observed with the BAT. However, it should be pointed out that performing the source control (i.e., catheter/stent removal, abscesses drainage) whenever possible as part of infection treatment might have contributed to the overall observed high clinical cure. Of note, the presence of bacteremic infections in onethird of subjects receiving the DC regimen strengthens the clinical effectiveness of this therapeutic option, which seems to retain its efficacy even in the presence of high bacterial inoculum, typically characteristic of bloodstream infections.
All the CP-CRE strains were K. pneumoniae: since KPC represents the most widely spread carbapenemase in our country, the results on the efficacy of DC might be translated even against KPC producing Enterobacteriaceae other than K. pneumoniae (i.e., E. coli) [23].
The results of the present study are in line with some recent investigations evaluating the clinical role of the DC regimen when no other options are available or after failure of first-line regimens [22] or in critically ill patients [27]. In the first study, the authors found a high clinical and microbiological success in a cohort of patients with complicated urinary tract infections (with or without secondary bacteremia), bloodstream infections, pneumonia, and external ventricular drainage infection [22] whereas in the second case-control study conducted in two Italian Intensive Care Units subjects receiving DC regimen presented with more severe clinical condition and had an improved 28-day mortality compared with those treated with standard regimen including colistin, tigecycline, or gentamicin [27].
Moreover, the efficacy of the DC regimen has been demonstrated in immunocompromised patients, including kidney transplanted patients [23,28] and a patient after allogenic hematopoietic stem cell transplantation [29].
One of the strengths of the present research is represented by the additional microbiological studies performed on the strains collected from patients receiving the DC regimen. In fact, automated systems such as VITEK-2, by indicating high carbapenems MIC as >16 g/mL, are unable to determine the precise MIC of carbapenems and there are growing evidences supporting the concept that knowing the real MIC of meropenem might influence the therapeutic choice and the effectiveness of carbapenem-based combination [15,30]. In particular, the DC appeared to be more effective in vitro if the meropenem MIC is ≤128 g/mL [15]. In the present study, the MICs 50/90 for meropenem and ertapenem were extremely high, with only 7/21 (33.3%) strains with meropenem MIC ≤128 g/mL; nevertheless, the overall clinical cure was as high as for strains exhibiting higher MICs. Interestingly, these results are similar to those found in a previous study [22] where the actual meropenem MIC, which was performed in 20/27 strains, was >256 g/mL in 5/20 (20%) strains, in the absence of clinical failure. Thus, the exact role of carbapenem MIC in predicting the DC clinical success should be better understood and deserves further investigations, together with the interaction with the patients' immune system.

Conclusions
In conclusion, we demonstrated that the DC regimen is a valid and effective therapeutic option in patients with infections due to KPC producing CRE, including those with bacteremic infection and more severe clinical conditions. The clinical effectiveness is maintained even in the presence of extremely high meropenem MIC.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare that they have no conflicts of interest.