What are the treatment options for resistant Klebsiella pneumoniae carbapenemase (KPC)-producing bacteria?

KEYWORDS:

• Antimicrobial stewardship
• Klebsiella pneumoniae carbapenemase
• gram-negative infections
• multidrug resistance

1. Introduction

In recent years, carbapenem-resistant Enterobacteriaceae (CRE) have been recognized as a common cause of healthcare-associated infections, causing high morbidity and mortality rates due to their difficult-to-treat resistance phenotypes and ability to rapidly disseminate in clinical settings [1].

Resistance to carbapenems in these organisms may be the result of a number of mechanisms, able to exert their action alone or in combination, including the production of serino β-lactamases (Klebsiella pneumoniae carbapenemase, or KPC, and OXA-48), metallo β-lactamases (Verona Integron-Mediated Metallo-β-lactamase, VIM, and New Delhi metallo-beta-lactamase, NDM) and ESBL (Extended-Spectrum Beta-Lactamases) and/or AmpC β-lactamases in association with alterations in the expression of outer membrane porins (OMPs) [2]. Adapted to persist in health-care settings and to colonize the intestinal tract, Klebsiella pneumoniae (Kp) is one of the most successful pathogen in causing nosocomial life-threatening infections [1,2]. Even though recently metallo-β-lactamase (MBL) producing Enterobacteriales have been increasingly observed among CREs that cause infections, Kp isolates producing KPC (KPC-Kp) remain predominant, with associated mortality rates ranging from 40% to 50% [3–5]. In more severely ill KPC-Kp infected patients, cornerstone of successful therapy is the administration of in vitro active antibiotics as early as possible [6]: thus, appropriate risk stratification for KPC-Kp infection and related mortality in the individual patient and/or fast microbiology results are crucial for prompt effective management [7,8]. In this editorial, we tried to summarize the treatment options for resistant KPC-producing bacteria, from what has been learnt from the past to what is the future going to bring us.

2. Available antimicrobial treatment options

Table 1 includes the last decade’s attempts made in order to assess an optimal therapeutic regimen for KPC-Kp infections, notwithstanding few available antimicrobial treatment options. Many in vitro active compounds such as colistin, tigecycline, and, more recently, ceftazidime/avibactam (CAZ-AVI) or meropenem/vaborbactam (MER-VAB) have been used, alone or combined with non-active in vitro drugs (i.e. meropenem) pursuing a synergistic effect [9–25]. All these drugs have been administered at high-doses (colistin 9 MU/day, tigecycline 200 mg/day, meropenem from 6 up to 11 gr/day) [4,9–25]. Prior to the approval of novel β-lactam/β-lactamase-inhibitor combinations (i.e. CAZ-AVI and MER-VAB), observational studies pointed out that combination therapy was associated with a better outcome, in terms of both clinical success and mortality rates [4,9–13]. In this regard, colistin-based combination therapy regimens have been predominately used and commonly considered the best available therapies (BATs), along with double carbapenem-based regimens (DCRs: ertapenem plus meropenem, alone or in combination with colistin or tigecycline). The postulated rationale of this combination was that ertapenem, whose activity is greatly affected by carbapenemases, may act as a suicide substrate, thus leading the other carbapenem to exert its antimicrobial activity [14–18].

Table 1. Clinical efficacy of old and new agents against KPC-Kp infections with comparison of monotherapy vs combination therapy

Reference	Study design	Enrolled patients	Infection	Microrganism	Regimen	Comparator regimen	Outcomes
Qureshi ZA et al. [9]	Retrospective observational two center cohort study	34 patients receiving definitive therapy (59% ICU patients)	BSIs	KPC-Kp	Monotherapies (n = 19: 37% colistin, 27% 21% tigecycline, 15% others)	Combination therapies (n = 15: 47% polimixins-based, 33% tigecycline based, 20% others)	28 day mortality rates were 58% and 13% with mono and combination therapies, respectively (P = 0.01) MVA: definitive therapy with combination therapies was associated with survival, P = 0.02.
Tumbarello M et al. [10]	Multicenter observational retrospective cohort study	125 patients (43% ICU patients)	BSIs	KPC-Kp	Monotherapies (n = 46: 48% colistin, 41% tigecycline or 11% gentamicin)	Combination therapies (n = 79: 79% tigecycline based, 66% colistin based. including double carbapenem based in 8 instances)	30 day mortality rates were 34% and 54% with mono and combination therapies, respectively; P =.02. MVA: postantibiogram therapy with tigecycline + colistin + meropenem was associated, P = 0.01
Daikos GL et al. [11]	Retrospective observational two center study	175 patients (58% ICU patients)	BSIs	CR-Kp (great majority KPC-Kp)	Monotherapies (n = 72: 51%, tigecycline, 30% colistin; 17% carbapenem, 12.5%aminoglycoside, 3% others)	Combination therapies (n = 103: 49% colistin based and 47% tigecycline based; 70% carbapenem-sparing regimens)	28 day mortality rates were 44% and 27% with mono and combination therapies, respectively, (P = 0.018).
Tumbarello M et al. [12]	Retrospective observational multicenter study	661 (n = 230 34.8% ICU patients)	Various (75.4% BSIs)	KPC-Kp	Monotherapies (n = 307: 39% colistin, 38% tigecycline, 23% gentamicin	Combination therapies (n = 354: 58% carbapenem based)	14 day mortality rates were 30% and 38% with mono and combination therapies, respectively, P = 0.03. MVA: Combination therapy was associated with lower mortality, P = 0.001
Falcone M et al. [4]	Retrospective observational study	111 (all ICU patients with septic shock)	Various (70% BSIs)	KPC-Kp	Monotherapies (definitive therapy with < two antibiotics displaying in vitro activity (n = 64)	Combination therapies (definitive therapy with ≥ 2 antibiotics displaying in vitro activity (n = 44)	30 day mortality rates were 85% and 15% with monotherapy and combination therapies, respectively (P < 0.001). MVA: colistin based regimen, use of ≥ 2 active agents as definite therapy and source control were associated with favorable outcome.
Souli M et al. [14]	Prospective observational study	27 (55% ICU patients)	Various (n = 12; 44% BSIs)	KPC-Kp	DC regimen (n = 27)	n/a	Clinical & microbiological success and 28 day mortality rates were 74% and 8.5%, respectively.
Cancelli F et al. [15]	Retrospective, single-center study	55 (no ICU patients)	Various (25.5% BSIs)	KPC-Kp	DC regimen: (n = 21)	BAT (mono/combination therapy with colistin, carbapenems, aminoglycosides, tigecycline or fluoroquinolones; n = 34)	60 day mortality rates were 9.5% and 11,7% with DC regimen and BAT, respectively
De Pascale G et al. Crit Care [16]	Case–control (1:2), observational two-center study	144 ICU patients)	Various (58% BSIs)	CR-Kp (n = 130; 90% KPC Kp; colistin resistant strains: 41.7% in the DC group vs 33.3%)	DC regimen (n = 48) plus colistin, and/or gentamicin and/or tigecycline in 73% of the cases.	Standard therapy with colistin or gentamicin or tigecycline (n = 96) alone (46%) or in combination (54%).	28-day mortality rates were 29% and 48% with DC and standard therapy, respectively (P = 0.04). MVA: DC regimen was associated with a reduction in 28-day mortality (P = 0.02)
Pea F et al [17]	Observational single center study	30 patients	Various (60% BSIs)	CR-Kp	Combination therapies including meropenem showing Css/MIC ratio <4 (regimen A, n = 16)	Combination therapies including meropenem showing Css/MIC ratio ≥4 (regimen B, n = 14)	Cures were 50% and 100% with regimen A and B, respectively (univariate logistic regression analysis P = 0.03)
Giannella M et al. [18]	Post hoc analysis of a multicenter observational study	595 (38% ICU patients)	BSIs	KPC-Kp	Combination therapies with HD meropenem (n = 428)	Combination therapies not including carbapenem (n = 167)	14 day mortality rates were 25% and 20% with HD meropenem and without carbapenem regimens, respectively, P = 0.18. When adjusted for the propensity score, HD meropenem-containing combination resulted protective, P = 0.03.
Shields RK et al. [20]	Retrospective, single center observational study	37 (29.7% transplant recipients)	Various	37 CRE; 28(75.7%) out of these were KPC-Kp)	CAZ-AVI alone (n = 26) or in combination with other antimicrobials (n = 11)	n/a	30-day clinical success: 59%. 30-day mortality rate: 24% 90-day recurrence: 23%. Microbiologic failure: 27% CAZ-AVI resistance: 10%
Shields RK et al [21]	Retrospective, single center observational study	109 (51.4% ICU patients)	BSIs	107 CR-Kp (97% KPC-Kp)	CAZ-AVI alone (n = 7) or in combination with gentamicin (n = 5)	Carbapenem plus aminoglycoside (n = 25), carbapenem plus colistin (n = 30), or other regimens (n = 41)	30-day clinical success and 90 day survival were higher among patients treated with CAZ-AVI than in those treated with all other regimens
Shields RK et al. [22]	Retrospective, single center observational study	77 (57.1% ICU patients)	Various	77 CRE (72.7% KPC-Kp)	CAZ-AVI alone (n = 53) or in combination other antimicrobials (n = 24)	n/a	30-day mortality rate: 19%. 90-day mortality rate: 31%. Clinical success rates were 36%, 75% and 88% for pneumonia, BSI and UTIs, respectively. At MVA, pneumonia and RRT were associated with clinical failure. Microbiologic failures: 32% (more common in pneumonia cases). RRT was an independent predictor of resistance development.
Van Duin D et al. [23]	Prospective, multicenter, observational study.	133 (59.1% ICU patients)	Various (46% BSIs)	137 CRE; (97% Kp). 52(96%) out of 54 Kp evaluated for presence or absence of carbapenemase genes were KPC-Kp.	CAZ-AVI alone (n = 24) or in combination with other antimicrobials (n = 14)	i.v. colistin (various dosing regimens) alone (n = 6) or in combination with other antimicrobials (n = 93)	All-cause 30 day mortality were 9% and 32% with CAZ-AVI and colistin based regimens, respectively (P = .001).
Guimarães T et al. [24]	Prospective multicenter observational study	29 (n = 17; 58.6% ICU patients)	Various (41.4% primary BSIs)	Enterobacteriales coresistant to carbapenems and polymyxins (100% KPC-2 producing strains)	CAZ-AVI with or without other antimicrobials	n/a	Overall clinical success: 82.7% (75% for BSIs) 14-day mortality: 31% 30-day mortality: 51.7% .
Tumbarello M et al [25]	Retrospective multicentre observational matched cohort control study (Italy)	208(32% ICU patients)	BSIs	All KPC-Kp	CAZ-AVI as salvage therapy alone (n = 22) or in combination to other antimicrobials (n = 82)	various regimems adopted as salvage therapy	30-day mortality rate were 36% and 56% with CAZ-AVI and other salvage regimens, respectively (P = 0.005) Receipt of CAZ-AVI was the sole independent predictor of survival in multivariate analysis with and without adjustment for propensity score.
Wunderink RG et al. [29]	Randomized (2:1), multicenter, multinational, open-label, active-controlled phase-3 trial	47(11% ICU patients; 25% immuno compromised patients)	Various (47% BSI)	47 confirmed/ suspected CRE (64% KPC-Kp)	MER-VAR (n = 32)	BAT (mono/combination therapy with polymyxins, carbapenems, aminoglycosides, tigecycline or CAZ-AVI alone; n = 15)	Clinical cure rates at EOT were 59% and 33% (P = 0.03) whereas 28 mortality rates were 15.6% and 33.3% (P = 0.20) with MER-VAB and BAT, respectively. Clinical failure or renal AEs rates were 28.1% and 80% with MER-VAB and BAT, respectively.
Ackley R et al et al. [30]	Retrospective multicenter observational study	130 patients (58% ICU patients)	Various (41% BSIs)	136 CRE (62% polimicrobial infections). KPC+CRE were isolated in approximately 70% of cases.	CAZ-AVI alone (n = 41) or in combination (n = 64) with other antibiotics (polimixins, n = 30)	MER-VAB alone (n = 21) or in combination (n = 4) with other antibiotics (polimixins, n = 2)	30-day mortality rates were 19% and 11.5% with CAZ-AVI and MER-VAB, respectively (P = 0.57). In the CAZ-AVI group, increase in MIC and emergence of resistance was observed 6 and 3 cases, respectively.
McKinnell JA et al. [35]	Phase 3, open-label study. Two cohorts: Cohort 1 (C1), a randomized comparison of PLZ vs. CST; Cohort 2 (C2), a single-arm study of PLZ in patients ineligible for C1 (CARE study)	39 patients (C1), 30 patients (C2)	Various (30/39, 76.9% BSIs in C1; 15/30, 50% BSIs in C2)	CRE (no attempt to define CR mechanism)	Cohort 1: PLZ (15 mg/kg q24 h) in combination with meropenem or tigecycline Cohort 2: PLZ plus investigator’s choice of adjunctive agent	CST (300-mg load [colistin base activity] then 5 mg/kg/d) plus tigecycline or meropenem.	All-cause mortality at Day 28 in patients with BSI from C1 was 7.1% (1/14) in PLZ group compared with 40% (6/15) in CST group. Higher rates of bacteremia clearance by Day 5 were observed in patients receiving PLZ as compared to CST (17/14, 85.7% versus 7/15 46.7%)

AEs, adverse events; BAT, best available therapy; BSIs bloodstream infections; CAZ-AVI, ceftazidime avibactam; CR, carbapenem resistance; CRE, Carbapenem-resistant Enterobacteriaceae; Css, steady-state concentration; DC regimen, double-carbapenem regimen; HD, high dose; ICU, Intensive Care Unit; KPC-Kp, Klebsiella pneumoniae carbapenemase–producing K. pneumoniae; LRTIs, low respiratory tract infections; MER-VAB, meropenem/vaborbactam; MIC, minimal inhibitory concentration; MVA, multivariate analysis; PDR, pandrug resistance phenotype; XDR, extensively resistance phenotype; RRT, Renal Replacement Therapy; UTIs, urinary tract infections; PLZ, plazomicin; CST, colistin.

Retrospective observational studies suggested that DCRs might be as effective as, or even superior to, BATs (which are usually colistin-based regimens) especially in the setting of infections caused by colistin-resistant strains [13,15,16]. Nevertheless, it should be pointed out that not only further controlled studies would be needed for a definite recommendation of DCRs but that their possible use must be very cautious given the relative epileptogenic potential and the risk of acquiring other carbapenem resistant pathogens (i.e. Acinetobacter baumannii) through the dysregulation of patient’s gut microbiota [19]. Anyhow, whatever are the agents used in combination with meropenem, the adoption of extended/continuous high doses infusion, capable of providing steady state serum concentration/MIC ratios ≥4, seems to be required for the optimal therapy [13,17,18].

CAZ-AVI has been the first β-lactam/β-lactamase-inhibitor combination available for the treatment of infections due to KPC-producing CRE, mainly represented by KPC-Kp. Until today, its use has been only supported by observational studies assessing its superiority as salvage therapy compared to BAT [20–25]. Notably, compared to colistin-based regimens, it has been advocated as superior in terms of clinical success, reduction of mortality rates and tolerability, both in monotherapy or combined with other antimicrobial agents [13,20–25]. In general, the potential usefulness of adding a second antibiotic to CAZ/AVI or MER-VAB has not been adequately investigated. Thus, no data showing a definite potential benefit of combination therapy with these agents are currently available. In particular, Tumbarello et al. showed that CAZ/AVI-based combination therapy, although with a favorable trend over CAZ/AVI alone, was not superior to monotherapy [25]. This finding has been further confirmed by a recent meta-analysis [26].

On the other hand, the alarming emergence of resistance in course of CAZ-AVI therapy has been described, possibly explained by conditions of pharmacokinetic (PK) underexposure as pneumonia or renal replacement therapy [20,22]. Different mechanisms have been associated with CAZ/AVI resistance: mutations in carbapenemases such as KPC and OXA-48, decreased permeability caused by mutations in outer membrane proteins and differences in susceptibilities of KPC subtypes. However, the most reported and identified resistance mechanisms in CRE, especially in Klebsiella pneumoniae and following CAZ/AVI treatment, are single amino acid substitutions in the KPC Ω-loop [27,28]. Under these circumstances, therapeutic drug monitoring may play a crucial role, since the synergistic effect of combination with other antimicrobial agents (i.e. meropenem, fosfomycin, or gentamicin) has not yet been demonstrated to be able to overcome PK underexposure [25,29].

CAZ-AVI is effective not only against KPC but also OXA-48-producing CRE [2], representing de facto the only viable therapeutic option for infections sustained by these strains.

MER-VAB, another beta-lactam/beta-lactamase-inhibitor combination active only against KPC-producing enterobacteriaceae with lack of activity toward OXA-48 and MBL, proved superior to BAT against KPC producing CRE infections in a prospective randomized study, conducted on a small cohort of subjects [29] and as effective as CAZ-AVI in a retrospective case series study [30].

Contrary to what has been described for CAZ/AVI, there is a lower propensity for developing resistance during therapy with MER-VAB [31–34]. In fact, resistance to MER-VAB in KPC-producing K. pneumoniae is rare, and, when it occurs, is a consequence of loss of function mutation in ompK35 and ompK36 or, as recently described, following a mutation of kvrA causing downregulation of ompK35/36 porin [32]. Furthermore, vaborbactam is less affected than avibactam by KPC-2 mutations conferring resistance to CAZ/AVI and, encouragingly, the risk of acquiring resistance during MER-VAB seems to be lower than that observed with CAZ/AVI [34]. Taken together, these data suggest that MER-VAB might represent the best available therapeutic option for KPC-producing CRE [30].

Plazomicin is a next-generation semisynthetic aminoglycoside active against MDR Enterobacteriales which received approval for c-UTI infections and showed improved outcomes compared with colistin in patients with bloodstream infections caused by CRE in the recent CARE trial [35,36]. Furthermore, its favorable lung penetration renders the drug a candidate for treatment regimens of VAP [35,36].

3. Antimicrobial agents on the pipeline

Table 2 summarizes novel compounds for the treatment of KPC-CRE infections. Among these, imipenem-relebactam, eravacycline (a tetracycline), and cefiderocol (the first siderophore cephalosporin) demonstrated good in vitro activity against KPC-CRE and received FDA approval for the indications reported in Table 2 [26–28]. However, clinical data of their efficacy against KPC-CRE are still poor or not currently available. Noteworthy, the efficacy of cefiderocol in the setting of carbapenem-resistant (CR) gram-negative infections has been questioned by the recent CREDIBLE-CR study, which overall showed higher rate of clinical failure in patients receiving cefiderocol versus BAT [37]. Nevertheless, these data mostly applied for non-fermenters (in particular for A. baumannii) whereas for carbapenem resistant K. pneumoniae the all-cause mortality at day 49 was similar between cefiderocol and BAT groups (23.5% versus 25.0%, respectively) [37].

Table 2. Summary of new agents with activity against KPC-Kp (references [31,38,39])

Agent	Class	CRE activity other than KPC	Other MDR activity	Route of administration	FDA approval	Potential uses or indications under investigation
Eravacycline	fluorocycline (tetracycline)	yes	A. baumannii, MRSA, VRE	Oral/i.v.	cIAIs	pneumonia, ABSSSIs
Apramycin	aminoglydoside	yes	A. baumannii, P. aeruginosa, MRSA	i.v.	-	-
Cefiderocol	siderophore cephalosporin	yes	A. baumannii, P. aeruginosa	i.v.	cUTIs	infections with limited treatment options
Cefepime /taniborbactam	β-lactam/β-lactamase-inhibitor	yes	P.aeruginosa	i.v	-	cUTIs, infections with limited treatment options
Cefepime /zidebactam	β-lactam/β-lactamase-inhibitor	yes	P. aeruginosa, A. baumannii	i.v	-	pneumonia infections with limited treatment options
Imipenem/ relebactam	β-lactam/β-lactamase-inhibitor	-	P. aeruginosa	i.v	cUTIs, cIAIs	infections with limited treatment options
Meropenem/ nacubactam	β-lactam/β-lactamase-inhibitor	Yes (OXA-48 producing strains only)	P. aeruginosa (hyper-AmpC-producing strains only)	i.v.	-	pneumonia. infections with limited treatment options
Meropenem/QPX7728	β-lactam/β-lactamase-inhibitor	yes	A. baumannii, P. aeruginosa	i.v.	-	-

A. baumannii, Acinetobacter baumannii; ABSSSIs, acute bacterial skin and skin structure infections; cIAIs, complicated intra-abdominal infections; CRE, carbapenem-resistant Enterobacteriaceae; cUTIs, complicated urinary tract infections; HABP/VAP, hospital-acquired/ventilator-associated bacterial pneumonia; I.v. intravenous; KPC-Kp, Klebsiella pneumoniae carbapenemase–producing K. pneumoniae; MDR, multidrug resistant; MRSA, methicillin resistant Staphylococcus aureus; P. aeruginosa, Pseudomonas aeruginosa; VRE, vancomicin resistant Enterococcus.

Except for apramycin (an aminoglycoside), all other agents on the pipeline are β-lactam/β-lactamase-inhibitor combinations [38,39]. Some of them have a broad spectrum of activity, including several multidrug resistant (MDR) Gram-negative bacteria, both CRE with carbapenem resistance mechanisms other than KPC production or non-fermenters MDR gram-negative isolates of Acinetobacter baumannii and Pseudomonas aeruginosa.

4. Expert opinion

At the present time and at least in the next future, KPC-CRE, mostly represented by K. pneumoniae, will remain a major public health issue, particularly in more debilitated patients as those critically ill or immunosuppressed. Current evidences pointed out that CAZ-AVI and MER-VAB provide a significantly better clinical outcome compared to old colistin-based regimens. These latter not only include agents with border-line in vitro activity against KPC-CRE activity but also are burdened with large between-patient variability and a substantial toxicity profile. Among the options in the pipeline, many other β-lactam/β-lactamase-inhibitors seem to show promising in vitro activity that warrants further study. It seems likely that in the next future these agents will represent the cornerstone of therapy for life-threatening infections caused by KPC-CRE. Hopefully, since almost all evidences of clinical efficacies of new agents against CRE are based on observational studies it is advisable that future investigation will be carried out with adoption of prospective randomized trials. At last, since emergence of resistance during therapy has been already observed with CAZ-AVI and MER-VAB [20,30,32], the use of all the new β-lactam/β-lactamase-inhibitors must be cautious and always within a well-defined antimicrobial stewardship program.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Additional information

Funding

This manuscript was not funded.