Carbapenem-resistant Enterobacteriaceae Infections: A Review of Epidemiology and Treatment Options

In 2013, the Centers for Disease Control (CDC) issued antimicrobial resistance guidance ranking carbapenemresistant Enterobacteriaceae (CRE), Neisseria gonorrhea, and Clostridium difficile as the three most urgent resistance threats in the United States [1]. CRE are defined as pathogens testing resistant to the following carbapenem antimicrobials (imipenem, meropenem, doripenem, or ertapenem) or are documented to produce a carbapenemase [2]. In the 2013 CDC report, an estimated 9,300 inpatient cases were predicted annually, and as of December 2017, CRE isolates have now been reported in all 50 states [3]. Enterobacteriaceae cause roughly 27.2% of healthcare-associated infections (HAIs) in acute care settings, with Klebsiella pneumoniae and E. coli as the predominant species [4]. Carbapenems are useful last line treatment options in multidrug-resistant gram-negative infections. Therefore, CRE are truly a healthcare threat.

Laboratory Standards Institute (CLSI) lowered the carbapenem breakpoints in an effort to facilitate detection and reporting of carbapenem-resistant isolates. The US Food and Drug Administration (FDA) shortly followed by updating medication package inserts. With the current guidance, supplemental testing outside of automated susceptibility testing instruments (AST) is no longer recommended. Institutions using older ASTs should update their equipment according to the newer breakpoints. Humphries R, et al. validated the accuracy of newer breakpoints to detect carbapenem-resistant isolates by comparing to polymerase chain reaction (PCR) assay for a variety of carbapenemase genes. Results indicated that few PCR-proven carbapenemase-producing isolates were reported as susceptible; <1% for ertapenem and imipenem, whereas 2.9% for meropenem [6]. Aside from AST, inaccuracies with the modified Hodge test have led to development of more reliable supplementary testing methodologies including the Carba-NP and carbapenem inactivation method. CLSI updates their testing recommendations annually, so please refer to that reference for the most up-to-date information.

CRE Infections
The 2013 CDC Urgent Threat Report identified the following patients at high risk for antibiotic resistant infections; immunocompromised patients (cancer chemotherapy, organ and bone marrow transplant, rheumatoid arthritis), dialysis patients, and complex surgical patients [1]. Closer examination of CRE infection literature confirms that dialysis and renal replacement therapy are indeed identifiable risk factors. Other risk factors include intensive care unit (ICU) admission, receipt of antimicrobials in the previous 90 days, cumulative antimicrobial exposures in the current admission, healthcare facility exposure/transfer, and poor functional status [9,10,11].
The impact of CRE infection on clinical outcomes is daunting. Using data from the Consortium on Resistance Against Carbapenems in Klebsiella and other Enterobacteriaceae (CRACKLE) study, a prospective, multicenter, observational cohort, Hauck et al. evaluated clinical outcomes in 260 CRE Klebsiella pneumoniae patients with bloodstream infection (BSI), pneumonia, and urinary tract infection (UTI) compared to those with CRE colonization and no active infection. A twofold increase in excess mortality was observed in patients with BSI and pneumonia (27% versus 12%), but not with UTI [12]. Other studies have described CRE BSI mortality rates between 40-70% [9]. A recent BSI study demonstrated that when carbapenem resistance was caused by a carbapenemase enzyme versus another resistance mechanism (ie, outer membrane porin changes), outcomes were worse. Authors suggest that carbapenemase-mediated resistance may be more virulent, although specific virulence factors were not elucidated [13]. An international, multicenter, retrospective study including 256 cases of CRE infection described overall 28-day mortality rates of 28.1%. A disparity was observed among severities of infection; 17.3% mortality with cUTI and pyelonephritis, but up to 44% with severe infections particularly with the following comorbidities were present; renal failure, sepsis, and immunedeficiency [14]. The mortality data discussed in this section was analyzed prior to the clinical use of newer agents that will be discussed below.
Also using data from CRACKLE, Eilertson et al. compared clinical outcomes in CRE-infected patients with healthy renal function versus those with renal replacement therapy (RRT). RRT patients are known from previous studies to have increased risk of mortality and infectious complications caused by multidrug-resistant pathogens including sepsis and pneumonia. In this study, those infected with CRE had worse outcomes than non-CRE RRT patients; 31% 28-day in hospital mortality [15]. Satlin et al. describe mortality rates of 40% in solid organ transplant and 65% of hematologic malignancy patients. Due to the potential for devastating patient outcomes in an immunocompromised population, the authors emphasize the importance of active surveillance and antimicrobial stewardship to aid in prevention of infection in these patients [16].

CRE Outbreaks
CRE isolates are a major focus of infection control programs since carbapenemases, especially KPCs, can be shared via mobile plasmids resulting in healthcare-associated institution outbreaks. It is important to implement effective infection control measures including contact isolation precautions, hand hygiene, surveillance, patient isolation, and environmental cleaning in patients with active or prior CRE infections [17]. In 2011, Detroit Medical Center reported an outbreak of colistinresistant NDM-producing Klebsiella pneumoniae involving three institutions caused by patient-to-patient transmission [18]. A rural Kentucky hospital reported a CRE outbreak occurring in 2016 from the emergency room to a specific hospital ward. During a 4-month period, the facility detected 23 CRE cases including both KPC and NDM1 carbapenemases. Through a CDC investigation, two observations were made; first, CRE isolates can be spread from urban to rural healthcare settings, and second, environmental cleaning carts (or any equipment moving throughout the facility) can be modes of transmission [19]. Outbreaks have been reported throughout the United States involving duodenoscopes or ventilators, invasive devices such as urinary catheters, hospital room sinks and bedrails, and have been spread across healthcare networks involving acute care and long-term acute care (LTAC) facilities [20,21].

CRE Treatment Options
CRE isolates are resistant to almost all beta-lactams leaving few, and unfortunately older, drug classes with adequate activity. Specifically, these older treatment options include aminoglycosides, polymyxins, a glycylcycline, and fosfomycin [22]. Table 1 provides dosing regimens and compares advantages and disadvantages of these older agents. It is important to note that polymyxin B and colistin (polymyxin E) are not identical in their adverse effect or pharmacokinetic profiles. At the time of publication, CLSI does not provide interpretive criteria for polymyxins for Enterobacteriaceae, and susceptibility testing for Pseudomonas aeruginosa is no longer recommended via Etest gradient diffusion or disk diffusion due to unreliable results [23]. Colistin methanesulfonate (CMS), a prodrug, requires renal elimination then hydrolysis for conversion to its active metabolite, colistin. This step is greatly impacted by renal dysfunction leading to high inter-and intra-patient variability in the serum concentrations of active drug. In patients with healthy renal function, 80% of CMS is eliminated in the urine before metabolism to the active drug even occurs, leaving subtherapeutic serum concentrations for non-urinary tract infections [24]. Polymyxin B is not impacted by this metabolic process and should be considered as the polymyxin of choice for systemic treatment of non-urinary tract infections. With regards to adverse effects, polymyxin B has also been shown to be less nephrotoxic to renal tubular cells [25]. A meta-analysis comparing differences in mortality and adverse effects between the two agents found no difference in clinical outcomes, but an increase in nephrotoxicity with colistin [26]. Future studies are needed to determine whether pharmacokinetic differences truly translate to undesirable patient clinical outcomes.
Alexander E, et al performed an international, multicenter CRE review using isolates from 2013-2014 and found overall antimicrobial resistance rates of greater than 80% to penicillins, cephalosporins, aztreonam, fluoroquinolones (FQs), and trimethoprim/sulfamethoxazole (TMP/SMX). The same CRE isolate review found antimicrobial resistance rates of 80% to tobramycin, 57.8% to amikacin, 43.1% to gentamicin, 37% to tigecycline, 26.4% to polymyxins [14]. The isolates in this study reflect an international summary of antimicrobial resistance and may not predict regional antimicrobial resistance trends. Antimicrobial stewardship programs (ASPs) have a profound effect on local institution resistance patterns by closely monitoring antimicrobial utilization. Luckily, newer betalactam agents and an aminoglycoside have been FDA approved in the United States to combat these highly resistant pathogens.
Treatment strategies utilizing these older agents is complex, and ongoing research regarding novel combination therapies has helped to optimize therapies. Combinations utilizing polymyxins, tigecycline, or aminoglycosides with or without a carbapenem have been described with mixed results. The polymyxins, aminoglycosides (except for treatment UTI), and intravenous fosfomycin should always be used in combination to prevent emergence of antimicrobial resistance. In a 2014 review of 20 studies using older agents to treat CRE infections, mortality rates were lower when combination therapy was utilized compared to monotherapy (27.4% vs. 38.7%; p < 0.001, respectively). In addition, combination regimens containing carbapenems resulted in the lowest mortality rates (18.8%) [27].
Other studies have found conflicting evidence regarding combination therapies with older agents. A 2014 meta-analysis of 692 patients described mortality rates for tigecycline-colistin and carbapenem-colistin regimens of 64% and 67%. Due to lack of statistically significant benefit of combination therapy in all infection types, authors concluded that it may only be beneficial in severely ill patients with severe infections [28]. Another review of 661 KPC infection episodes in Italy illustrated that only those patients with high-risk infection (defined as non-UTI) or a meropenem MIC ≤8 mg/L. benefited from combination therapy [29]. In a recent randomized, controlled trial of 406 patients with pneumonia or bacteremia, Paul M et al evaluated colistin alone versus colistin plus meropenem for carbapenem-resistant gram-negative infections [30]. Although the majority of infections in this study were caused by Acinetobacter baumanii instead of CRE isolates, combination therapy was not superior to monotherapy and resulted in an increase in adverse effects, most notably diarrhea.

Newer Agents with CRE-Activity
Ceftazidime-avibactam (CAZ-AVI) is an intravenous cephalosporin/beta-lactamase inhibitor combination. Avibactam, a novel non-beta-lactam beta-lactamase inhibitor (NB-BLI) diazabicyclooctane (DBO), is currently commercially available as a combination agent with ceftazidime, and is also being studied in combination with aztreonam and ceftaroline [34]. CAZ-AVI inhibits Class A, C, and some D enzymes, but is inactive against Class B MBLs. It is most clinically useful for its activity against extended-spectrum beta-lactamases (ESBLs) and KPC carbapenemases. In 2015, CAZ-AVI was FDA approved for treatment of complicated urinary tract infections including acute pyelonephritis (cUTI/AP) and complicated intra-abdominal infections (cIAI) in addition to metronidazole. In early 2018, CAZ-AVI also gained FDA approval for hospitalacquired pneumonia (HAP) and ventilator-associated pneumonia (VAP). The pneumonia trial comparator was meropenem 1g IV q8h [35]. CAZ-AVI should be administered 2.5g IV q8h infused over 2 hours, and a renal dosage adjustment is required for patients with a creatinine clearance (CrCl) <50 ml/min. Reported adverse effects from Phase 3 trials are similar to ceftazidime monotherapy.
Although patients infected with CRE-producing strains were excluded from the cUTI, cIAI and pneumonia trials due to lack of CRE activity with the comparator agents, newer clinical experiences describe efficacy and use of combination therapy for CRE infections. Two retrospective studies evaluated clinical success and mortality rates in CRE patients treated with CAZ-AVI. In both studies, Klebsiella pneumoniae was the predominant strain. Shields RK, et al. evaluated 30-day clinical success and mortality rates in thirty-seven cases. Fifty percent were diagnosed with HAP or VAP, and 70% were treated with monotherapy [36]. King et al, evaluated end of treatment clinical success and mortality rates in sixty cases [37]. The most common infection types were bacteremia, pneumonia, and UTI, and 55% were treated with monotherapy. Between the two studies, clinical success rates were similar regardless of the timepoint, 59% and 65% respectively. Mortality rates were also similar; in-hospital mortality 32%, and 30-and 90-day mortality, 24% and 38%. CAZ-AVI resistance developed during therapy in both studies. The King study found no statistically significant difference in outcomes between monotherapy and combination therapy.
Other studies have compared CAZ-AVI to alternative agents for treatment of CRE infections. A comparative study between CAZ-AVI-containing regimens and alternative regimens (mostly combinations with carbapenems) in patients with KPC-Klebsiella pneumoniae bacteremia demonstrated higher clinical success rates (85% versus 37-48%, P = 0.04). 90-Day mortality rates were lowest among patients receiving CAZ-AVI regimens.
Patients receiving combination regimens containing an aminoglycoside or colistin experienced a higher rate of nephrotoxicity [38]. Another comparative study (n = 137 patients), using data from the CRACKLE database, analyzed efficacy, safety, and benefit-risk outcomes between CAZ-AVIand colistin-containing initial treatment regimens. Patients had a greater probability of a better outcome in all categories with initial treatment with CAZ-AVI versus colistin. Fourteen (37%) patients were treated with monotherapy in the CAZ-AVI arm [39]. Meropenem-vaborbactam (MVB) is an intravenous carbapenem/ beta-lactamase inhibitor combination. Vaborbactam is the first non-beta-lactam boronic acid inhibitor with activity against Class A and C beta-lactamases, and is inactive against Class B MBLs. Similar to CAZ-AVI, MVB is also most clinically useful for its activity against ESBLs and KPC carbapenemases [40]. In 2017, MVB was FDA approved for treatment of cUTI/AP (TANGO-1 trial). The comparator was piperacillin-tazobactam, and patients could be switched to oral levofloxacin to complete treatment. Few CRE infections were included in this trial [41].
MVB 2g/2g IV should be administered every 8 hours infused over 3 hours, and a renal dosage adjustment is required for patients with a creatinine clearance (CrCl) <50 ml/min. Reported adverse effects from Phase 3 trials are similar to meropenem monotherapy.
The TANGO-2 trial is a Phase 3 study including patients with CRE infection (59.7%). Patients with cUTI/pyelonephritis, cIAI, bacteremia, nosocomial pneumonia were included, and the comparator was "best available therapy" (BAT), meaning one of the older anti-CRE agents discussed above as monotherapy or in combination. Recognizing that CRE patients have multiple comorbidities and are different than typical study populations for these types of infections, study investigators performed a retrospective review of patients with CRE infections in order to better define study design enrollment criteria [14]. Authors estimate that only 22% of the TANGO-2 study population would have met enrollment using traditional criteria. As a result of the study, it is unique that the TANGO-2 study allowed immunocompromised patients, and those with severe sepsis and renal insufficiency needing dialysis, although these types of patients did not comprise the majority of those enrolled. Enrollment in the BAT arm was discontinued early when interim analysis revealed better outcomes in MVB-treated patients; in CRE patients, clinical cure 57.1% versus 26.7%. Despite promising clinical outcomes, minimum inhibitory concentrations (MIC) increased in several MVB-treated isolates [42,43].
Plazomicin (PZM) is an intravenous aminoglycoside antimicrobial with activity against ESBL-and carbapenemaseproducing aerobic gram negative rods. It is stable against aminoglycoside modifying enzymes that inactivate gentamicin, tobramycin, and amikacin, and most carbapenemase classes, including some activity against Class B MBLs (except NDM) [44]. There is limited activity against Stenotrophomonas maltophilia, Acinetobacter baumanii, and Pseudomonas aeruginosa, and future studies are needed to determine clinical utility for these pathogens.
In 2018, PZM gained FDA approval for treatment of cUTI/AP (EPIC trial) [45]. The comparator was meropenem 1g IV q8h as a 30 minute infusion, and a switch to an oral agent was allowed at 4 days. PZM was studied using extended interval dosing, 15 mg/kg IV once daily over 30 minutes, and a renal dosage adjustment is required for patients with a creatinine clearance (CrCl) <60 ml/min. Forty percent of Enterobacteriaceae isolates were an ESBL-producing or aminoglycoside non-susceptible strain. Therapeutic drug monitoring of a serum trough concentration (goal ≤3 mg/L) is recommended 30 minutes prior to the second dose.
With regards to adverse effects, in the EPIC study the incidence of nephrotoxicity in the PZM arm, defined as a serum creatinine increase of ≥0.5 mg/dL, was 3.6%. Results from another Phase 2 study comparing PZM to levofloxacin 750 daily for cUTI/AP suggest higher nephrotoxicity when 15 mg/ kg was used compared to 10 mg/kg, although the higher dose was selected for Phase 3 trials to optimize pharmacokinetic and pharmacodynamic targets [46]. A Phase 3, open label, Combating Antibiotic-Resistant Enterobacteriaceae (CARE) study evaluated plazomicin versus colistin in combination with meropenem or tigecycline for serious CRE infections (BSI and HABP/VABP). Results include lower 28-day all-cause mortality, higher microbiological response rates, and reduced nephrotoxicity. Authors concluded that plazomicin-based combination therapy had a better efficacy and safety profile compared to colistin-based combination regimens [47,48]. Although classwide warnings for nephrotoxicity, ototoxicity, neuromuscular blockage, and fetal harm were included in product labeling, the incidence of nephrotoxicity appears to be reduced compared to colistin or earlier aminoglycosides, especially when therapeutic drug monitoring is utilized.
In summary, for the two newer beta-lactam agents CAZ-AVI and MVB, safety and efficacy data appear favorable compared to older therapies for CRE infections. It Is concerning that MIC elevations, some conferring resistance to both CAZ-AVI and MVB, were observed. PZM offers a potential expanded CRE spectrum of activity and a better safety and efficacy profile compared to colistin-based combination regimens. Future studies are needed to optimize protection against antimicrobial resistance through combination therapy or optimal dosing.'

Antimicrobial Pipeline
Several agents with novel mechanisms of action are in the antimicrobial development pipeline. Table 2 lists five agents with Phase 2 and 3 clinical trial data, and likely review for FDA approval in the near future. Both non-CRE and CRE studies were included in the table.
Conclusions CRE infections are a healthcare and public health threat. Mortality rates are high, and agents with better safety and efficacy profiles plus expanded carbapenemase activity are desperately needed. When using older agents, combination therapy, particularly carbapenem-containing regimens, reduces mortality with severe infections, but may not have as much of a benefit with UTIs. Three newer agents, CAZ-AVI, MVB and PZM, are available and have activity against Class A-producing Klebsiella pneumoniae, the predominant CRE pathogen in the United States. In small studies, these agents demonstrate promising these clinical outcome and mortality data compared to older agents. However, further studies are needed to optimize their use (monotherapy versus combination therapy) and prevent emergence of therapy. Infections caused by MBLs remain infrequent in the United States, and treatment options against these strains are sparse. Several antimicrobial agents are in the pipeline to address this need.