Management of infections due to KPC-producing Klebsiella pneumoniae.

The emergence of the Klebsiella pneumoniae carbapenemases in K. pneumoniae and other Gram-negative bacteria, usually on a background of multidrug resistance, has led to difficult therapeutic choices. Among available antibiotics, tigecycline and the polymyxins are the most frequently active against these organisms in vitro. Optimal therapy of infections due to these bacteria may involve maximization of antibiotic dose as well as their use in combination.

Introduction and context

A novel class A serine carbapenemase detected in North Carolina in 1996 was given the designation KPC-1 because it was first identified in Klebsiella pneumoniae [1]. Since then, a number of variants of this enzyme (currently up to KPC-8) have been identified and have spread globally [2]. The mobile nature of the genetic element encoding KPC, Tn4401, which is carried on a plasmid, has contributed to the spread of this enzyme, which has now been identified in many of the Enterobacteriaceae as well as in Pseudomonas and Salmonella [2]. Strain dissemination, however, appears to be of greater importance than plasmid dissemination as evidenced by the fact that almost 70% of strains in a US Centers for Disease Control and Prevention collection are of a single lineage, sequence type (ST) 258 [3].

KPCs hydrolyze all β-lactam antibiotics and are not significantly inhibited by clavulanic acid, sulbactam, or tazobactam [2]. Detection of KPCs in the clinical laboratory may be difficult and many bacteria producing this enzyme appear to be susceptible to anti-pseudomonal carbapenems by routine testing [4]. Resistance to ertapenem may be a more sensitive (albeit less specific) indicator of the possible presence of KPC than resistance to other carbapenems. Because of the insensitivity of current breakpoints in detecting the presence of KPC, it is suggested that isolates of K. pneumoniae with a minimal inhibitory concentration (MIC) of at least 1 μg/mL to imipenem, meropenem, or doripenem undergo further testing designed to detect the presence of carbapenemases, including KPC [4-6]. The importance of detection resistance in these organisms is evidenced by the fact that the use of imipenem or meropenem in patients with infections due to KPC-producing K. pneumoniae that appear to be susceptible to these carbapenems is associated with frequent therapeutic failure [7].

KPCs are often accompanied by additional β-lactamases within individual isolates [4]. A survey of 42 KPC-producing K. pneumoniae isolates in the Eastern US found that each isolate carried a mean of 3.5 β-lactamases [4]; one KPC-producing K. pneumoniae isolate has been reported to contain seven extended spectrum β-lactamases [8]. K. pneumoniae isolates producing both KPC-2 and VIM-1 enzymes (the latter of which are class B metallo-carbapenemases) have been detected in Greece [9].

Recent advances

KPC-producing bacteria are often resistant to a broad range of antibiotics beyond the β-lactams. In fact, the plasmids that carry the KPC gene may also contain genes encoding aminoglycoside-modifying enzymes and some have also been reported to encode QnrA and QnrB, resulting in reduced susceptibility to fluoroquinolones [4,10]. An evaluation of 104 carbapenemase-producing Enterobacteriaceae, 70% of which produced KPC-2 or KPC-3, found that the most active agent tested was tigecycline, to which all isolates were susceptible [11]. Among antibiotics other than β-lactams that were tested, tigecycline was followed (in decreasing order of susceptibility) by polymyxin B (88% susceptible), amikacin (73.0%), gentamicin (50.0%), tetracycline (35.6%), and ciprofloxacin (32.1%). A similar evaluation of 96 carbapenemase-resistant K. pneumoniae from 10 Brooklyn hospitals (82% of a single ribotype) also found that all were susceptible to tigecycline [12]. Ninety-one percent were considered susceptible to polymyxin B, 66% to doxycycline, 61% to gentamicin, 45% to amikacin, and 2-7% to chloramphenicol, rifampin, tobramycin, and ciprofloxacin. Time-kill studies found that tigecycline was bacteriostatic whereas gentamicin and polymyxin B were bactericidal. Killing by polymyxin B was concentration-dependent.

Most published experience of treatment of infection due to KPC-producing organisms is with tigecycline, a polymyxin (polymyxin B or colistin), or their combination, with variable reported outcomes. Monotherapy, however, may be associated with persistent infection and the emergence of antibiotic resistance. Recurrence of an empyema due to a KPC-producing K. pneumoniae during treatment with tigecycline was associated with an increase in MIC from 0.75 to 2.0 μg/mL [13]. Elemam and colleagues [14] recently reported two patients with infections due to K. pneumoniae isolates resistant to all antibiotics tested, including tigecycline and polymyxin B, with both isolates having developed progressive increases in the MIC to polymyxin B during treatment with this antibiotic. Similarly, of 16 patients with infection due to carbapenemase-resistant K. pneumoniae carrying a KPC gene who had persistently positive cultures, 12 were treated with polymyxin B alone whereas 4 received this polypeptide antibiotic in combination with tigecycline for at least a portion of their therapy [15]. Blood cultures, though initially positive in only 12, subsequently yielded the organism in all 16. Marked increases in the polymyxin B MIC were observed in 3 of the 12 given that antibiotic alone and in 0 of 4 who also received tigecycline. The development of resistance during therapy may be due to the presence of heteroresistant subpopulations, a phenomenon observed with colistin (polymyxin E) in 15 of 16 multidrug-resistant K. pneumoniae isolates considered susceptible to colistin by MIC testing, a result consistent with the very high mutant prevention concentration/MIC ratios observed [16]. Although polymyxin B causes concentration-dependent killing, this is accompanied by rapid regrowth of resistant subpopulations with little or no postantibiotic effect [16], consistent with the clinical observation of treatment-emergent resistance. It should be noted that resistance has also been reported to emerge in Acinetobacter during treatment with tigecycline [17].

This problem may require optimization of dosage regimens to overcome some of these characteristics of polymyxins, combination therapy, or both. Very limited published clinical experience suggests that combination therapy may prove to be more successful than polymyxin monotherapy. Neither antagonism nor synergy was detected with the combination of tigecycline and polymyxin B tested against three isolates of K. pneumoniae in vitro [18]. In one series, only 1 of 3 patients with bacteremia due to a KPC-2-producing K. pneumoniae had a favorable outcome after treatment with polymyxin B, whereas another patient, treated with a combination of tetracycline and an aminoglycoside, survived [19]. As indicated above, resistance to polymyxin B occurred in 0 of 4 patients given this antibiotic in combination with tigecycline, whereas resistance occurred in 3 of 12 given polymyxin B monotherapy [15]. Other combinations, such as those including an active aminoglycoside, may also have promise: a patient with endocarditis due to KPC-3-producing K. pneumoniae was successfully treated with a combination of colistin and gentamicin [20]. Another agent for consideration in combination therapy is rifampin; polymyxin B and rifampin were synergistic in vitro against 15 of 16 isolates of carbapenemase-resistant K. pneumoniae [12].

Implications for clinical practice

The optimal therapy for infections due to these multidrug-resistant pathogens is not well defined and depends upon the susceptibilities of individual isolates, and the choices are often severely limited. The most frequently active antibiotics in vitro are tigecycline and colistin (or polymyxin B) and these may often be the only choices based on susceptibility test results. The pharmacodynamics of the polymyxins remain poorly defined, and optimal dosing regimens have not yet been determined [21]. It is possible that doses higher than those currently approved by the US Food and Drug Administration may provide improved therapeutic results. Concerns about toxicity of the polymyxins (which may have previously been overestimated) have limited the attempts at significantly increasing the dose of these agents. In cases of ventilator-associated pneumonia, consideration could be given to adjunctive administration of colistin by aerosolization [22].

The currently approved dose of tigecycline results in low serum and urine concentrations, raising concern about the efficacy of this agent in the treatment of bacteremia and urinary tract infections as well as about the emergence of resistance [17]. Inadequate tissue concentrations may have accounted for the failure of tigecycline relative to comparator therapy in the treatment of patients with ventilator-associated pneumonia due to a variety of organisms [23]. Cunha [24] has reported administering tigecycline in a 400 mg loading dose followed by 200 mg every 24 hours - a dose four times higher than currently approved - without eliciting nausea or vomiting, frequent adverse reactions seen with standard dosing regimens. However, he provides little specific information in regard to the use of these doses.

Although higher doses of tigecycline or a polymyxin may be beneficial, lack of knowledge in regard to tolerability and safety may limit their current acceptance by clinicians. Combination therapy of serious infections due to KPC-producing organisms would, however, appear to be warranted. The combination to be used depends on the antibiotic susceptibilities of the isolated pathogen but most often would be tigecycline with a polymyxin, perhaps at escalated dosages. Limited in vitro data suggest that co-administration of rifampin may have a role in such cases. Repeated cultures during therapy should be performed with monitoring for emergence of resistance.

The future of therapy for infections caused by multidrug-resistant aerobic Gram-negative bacilli appears bleak as a result of the paucity of development of new drugs targeting these organisms [25]. One potential exception is the development of β-lactamase inhibitors active against carbapenemases, such as NXL104, a potent inhibitor of KPC-2 [26]. It can be hoped that the attitude of the pharmaceutical industry toward antibiotic development will soon change and that additional candidates will be developed.