Since the first description of extended-spectrum β-lactamase (ESBL) production by
Klebsiella pneumoniae in 1983 (
11), antibiotic-resistant strains that produce ESBLs have emerged among the members of the family
Enterobacteriaceae, predominantly in
Escherichia coli and
K. pneumoniae, and isolates resistant to broad-spectrum cephalosporins are increasingly being recognized (
3). In the past decade, a new problem has emerged in enteric bacteria: plasmid-mediated AmpC enzymes. They are derived from chromosomal AmpC genes of gram-negative organisms, such as
Citrobacter freundii,
Enterobacter cloacae, and
Aeromonas species (
21).
Organisms with plasmid-mediated AmpC enzymes are generally resistant to broad-spectrum penicillins, extended-spectrum cephalosporins, monobactam, and cephamycins but are susceptible to cefepime, cefpirome, and carbapenems (
21). However, it is difficult to distinguish ESBL-producing organisms from plasmid-mediated AmpC β-lactamase-producing organisms by phenotypic susceptibility testing. Standard guidelines for the detection of AmpC-producing isolates are also lacking.
In this report, we describe the epidemiology and microbiological characteristics of AmpC β-lactamase-producing K. pneumoniae isolates and analyze the clinical characteristics of the patients infected by AmpC enzyme-producing K. pneumoniae isolates. In addition, we compared the clinical features and outcomes of bloodstream infections caused by AmpC β-lactamase-producing K. pneumoniae isolates with those caused by TEM- or SHV-related ESBL-producing K. pneumoniae isolates.
DISCUSSION
Since plasmid-mediated AmpC β-lactamase was first reported from
K. pneumoniae in 1989 (
2), plasmid-mediated AmpC β-lactamases have increasingly been identified worldwide. Plasmid-mediated AmpC β-lactamases have been discovered most frequently in
K. pneumoniae isolates and also in other species naturally negative for AmpC, such as
Klebsiella oxytoca,
Salmonella, and
Proteus mirabilis (
21).
In our study, on the basis of an analysis of the cases of
K. pneumoniae bacteremia detected at a single institute in South Korea, DHA-1- and CMY-1-producing isolates were found to be common among the isolates resistant to extended-spectrum cephalosporins. Previous studies showed that CMY-1 is prevalent in Korea (
8,
9,
19); however, it should be noted that DHA-1, an inducible AmpC β-lactamase, is prevalent at the Seoul National University Hospital, a 1,500-bed university hospital. Since the first description of DHA-1 from a strain in Saudi Arabia in 1998 (
1), DHA-1-producing clinical isolates have been reported in Taiwan (
29).
The DHA-1 enzyme, which is mediated by 110-kb plasmid, was first identified from
K. pneumoniae strain 502321 in Korea in 2000 (unpublished data). We cloned and sequenced nucleotides of the gene and found that the sequence of the
bla gene of this isolate was identical to that of
blaDHA-1. An
E. coli DH10B isolate containing this clone showed a resistance pattern identical to that of
E. coli HB101(pSAL-1) (
28): resistance to streptomycin and sulfonamides.
It is noteworthy that several geographic clusters of AmpC β-lactamase types have been described. These include a North American cluster (MIR-1 and ACT-1), a Central and South American cluster (FOX-1 and FOX-2), an Asian cluster (CMY-1 and MOX-1), and a Mediterranean and Middle Eastern cluster (CMY-2, CMY-2b, LAT-1, and LAT-2) (
21). Because few laboratories test for the production of the AmpC β-lactamase and even fewer laboratories test for induction, the occurrence of these enzymes in
K. pneumoniae and
E. coli isolates remains uncertain, as do their impacts on therapies and clinical outcomes.
In this study, we evaluated the clinical features and outcomes of bloodstream infections caused by AmpC-type β-lactamase-producing
K. pneumoniae isolates. In addition, these patients were compared to those infected with ESBL-producing
K. pneumoniae isolates. The clinical characteristics were similar to those caused by TEM- or SHV-related ESBL producers. Previous studies demonstrated that prior use of antibiotics (
9,
12), the presence of a central venous catheter or a urinary catheter (
22), and prior hospitalization and the use of extended-spectrum cephalosporins (
9) are risk factors for infections caused by ESBL-producing
K. pneumoniae or
E. coli isolates.
Analysis of the clinical outcomes demonstrated high rates of failure of the initial antimicrobial therapy, especially cephalosporin treatment, in patients infected with AmpC β-lactamase-producing organisms, as was the case for patients infected with TEM- or SHV-related ESBL producers. Although the number of patients was small and the patients were not controlled for the severity of disease, the 30-day mortality rate was higher in the DHA-1 group than in the CMY-1-like group (46 and 14.3%, respectively). The mortality rate for the patients who received extended-spectrum cephalosporins as definitive treatment was assessed: all four patients in the DHA-1 group died, and one of three patients in the CMY-1-like group died. This result might be partially explained by the fact that β-lactamases had been induced by exposure to β-lactam antimicrobials in DHA-1-producing isolates, thus providing higher levels of resistance.
In the present study, all but three AmpC β-lactamase-producing isolates (one CMY-1 producer and two DHA-1 producers) were susceptible to cefepime. These results suggest that cefepime might be useful for the treatment of infections caused by AmpC β-lactamase-producing organisms (
29). However, a report (
15) has described the inoculum effect of cefepime or cefpirome in an AmpC producer, which lacked an outer membrane protein. In our study, all the patients treated with extended-spectrum cephalosporins received cefotaxime or ceftazidime, but not cefepime, since cefepime was not available at the Clinical Research Institute of Seoul National University Hospital until recently. Nevertheless, further studies to determine whether cefepime can be used for the treatment of infections caused by plasmid-mediated AmpC β-lactamase producers are needed.
It is difficult to distinguish organisms producing ESBLs from those producing plasmid-mediated AmpC β-lactamases by phenotypic susceptibility testing. Resistance to cefoxitin indicates the possibility of AmpC-mediated resistance but also indicates reduced outer membrane permeability. Some phenotypic tests are available to help distinguish the difference between cefoxitin-resistant non-AmpC producers and cefoxitin-resistant AmpC producers. These include a three-dimensional test (
26) and a new AmpC disk test (J. A. Black, E. S. Moland, and K. S. Thomson, Abstr. 42nd Intersci. Conf. Antimicrob. Agents Chemother., abstr. D-534, 2002). In addition, the use of β-lactamase inhibitors can help identify possible AmpC-producing organisms (
25). However, none of these tests are standardized and they are time-consuming, especially for a clinical microbiology laboratory handling large numbers of isolates.
Reporting of a susceptibility testing result for AmpC β-lactamase producers can be controversial if they show susceptibility to some extended-spectrum cephalosporins in vitro, because no standard method for the detection of these isolates is yet available. Moreover, there are few clinical data on the patients infected with these organisms. Although the number of patients in our study was small, the study has shown that the outcome of cephalosporin treatment for serious infections due to AmpC β-lactamase-producing K. pneumoniae isolates was poor, even for infections caused by apparently susceptible organisms. Therefore, a standard test for the detection of the plasmid-mediated AmpC enzyme and new breakpoints for extended-spectrum cephalosporins are urgently necessary.
To the best of our knowledge, this is the first description of the clinical features and outcomes of bloodstream infections caused by AmpC β-lactamase-producing K. pneumoniae isolates.