Surveillance of Community-Based Reservoirs Reveals the Presence of CTX-M, Imported AmpC, and OXA-30 β-Lactamases in Urine Isolates of Klebsiella pneumoniae and Escherichia coli in a U.S. Community

Abstract

β-Lactamases produced by urine isolates from patients in long-term care facilities (LTCFs), outpatient, clinics, and one hospital in a U.S. community were characterized. A total of 1.3% of all Escherichia coli and Klebsiella pneumoniae isolates collected from patients in 30 LTCFs and various outpatient clinics produced extended-spectrum β-lactamases (ESBLs) and/or imported AmpC β-lactamases.

Pathogens producing extended-spectrum β-lactamases (ESBLs) and imported AmpC β-lactamases have become endemic in some hospitals, limiting therapeutic options for antibiotic treatment (3, 4, 7). Reports from other countries indicate that patients from long-term care facilities (LTCFs) and community-based sources are infected with gram-negative pathogens producing ESBLs or imported AmpC β-lactamases (6, 15-18, 20). However the evidence describing this potential problem in the United States is minimal and outdated (2, 5, 22, 24). Nursing homes and other community-based patient sources of imported AmpC β-lactamases and ESBLs can serve as reservoirs for strains causing infections in the hospital setting (24). A recent study of the trends of antibiotic resistance in U.S. nursing homes has documented that nearly 1,200 new cases of antibiotic-resistant bacterial infections occur in U.S. nursing homes every year (10). In addition, animal reservoirs harboring ESBLs or AmpC-producing bacteria have been identified (8, 21).

Community-based sources of plasmid-encoded ESBL- and AmpC-mediated resistance can pose a serious risk of transmission to hospitalized patients when infected or colonized patients are admitted, presenting a concern for hospital infection control (24). Surveillance of these resistance mechanisms is thus warranted. Therefore, to evaluate the potential for community-acquired ESBLs and imported AmpC β-lactamases, a study was designed to characterize the β-lactamases produced by urine isolates from patients in nursing homes, outpatient clinics, and one hospital in Omaha, NE.

Between 1 January and 30 June 2005, 1,433 Escherichia coli and 214 Klebsiella sp. consecutive urine isolates from patients in 30 nursing homes, various outpatient clinics, and Creighton University Medical Center (CUMC) were screened in the Microbiology Laboratory of CUMC using disk diffusion as recommended by the Clinical and Laboratory Standards Institute (CLSI) (9). Isolates that were nonsusceptible to cefoxitin and/or cefpodoxime (NS-FOX/CPD) were collected. Seventy-five E. coli and 14 Klebsiella isolates met these criteria.

The presence of ESBLs was determined using the ESBL confirmatory disk test according to CLSI recommendations and isoelectric focusing (IEF) with inhibitor overlays as previously described (23). Broth microdilution was performed using Trek Diagnostic Systems (Cleveland, OH) panels and interpreted according to CLSI recommendations. The panel contained ceftazidime and ceftazidime in combination with clavulanate, using a range of 0.06 to 128 μg/ml, and cefotaxime and cefotaxime in combination with clavulanate, using a range of 0.06 to 64 μg/ml. A set concentration of 4 μg/ml was used for clavulanic acid. AmpC production was examined using the AmpC disk test (1). To identify the production of an ESBL in the presence of an AmpC β-lactamase, 200 μg/ml of cloxacillin (to inhibit the AmpC β-lactamase) was added to Mueller-Hinton agar when the ESBL confirmatory disk test was performed. Identification of the genes encoding the ESBL or the AmpC type was done by PCR and sequence analyses as previously described (12, 14, 19). Sequence alignment and analysis were performed online using the BLAST program of the National Center for Biotechnology Information (www.ncbi.nlm.nih.gov).

Seven isolates (CUMC243, -181, -190, -214, -223, -224, and -225) were false positive and one isolate (CUMC215) was false negative for ESBL production by CLSI disk diffusion confirmatory testing (Table 1). All eight discrepancies were resolved by adding 200 μg/ml of cloxacillin to the medium. Interestingly, all of the ESBL false-positive isolates tested negative for ESBL production using microdilution (data not shown). All imported AmpC-producing organisms were resistant to cefoxitin and tested positive with the AmpC disk test. Isolates producing ESBL without the coproduction of AmpC were susceptible to cefoxitin and tested negative with the AmpC disk test.

TABLE 1.

Susceptibility and molecular results for ESBL- and/or AmpC-producing E. coli and K. pneumoniae isolates

Isolate	Species	Susceptibilitya to:				pI(s)b	PCR/sequencingc	Sited	β-Lactamase(s) producede
		CAZ	CAZ-CA	CTX	CTX-CA
CUMCK2f	K. pneumoniae	8 (R)	12 (−)	11 (R)	16 (+)	9.2, 8.0, 5.6	CMY-2, CTX-M-14, TEM-like	NH1 (patient 1)	CMY-2, CTX-M-14, TEM-2-like
CUMC215	E. coli	6 (R)	7 (−)	7 (R)	8 (−)	9.2, 8.0, 5.4	CMY-2, CTXM-14, TEM-like	NH1 (patient 2)	CMY-2, CTX-M-14, TEM-1-like
CUMC243g	E. coli	7 (R)	12 (+)	11 (R)	12 (−)	9.2	CMY-2	NH2	CMY-2
CUMC181	E. coli	6 (R)	14 (+)	13 (R)	13 (−)	9.2	CMY-2	NH3	CMY-2
CUMC190	E. coli	8 (R)	13 (+)	6 (R)	11 (+)	9.2	CMY-2	NH3	CMY-2
CUMC201	E. coli	8 (R)	12 (−)	12 (R)	10 (−)	9.2	CMY-2	IP1	CMY-2
CUMC214	E. coli	8 (R)	13 (+)	13 (R)	12 (−)	9.2	CMY-2	OP4	CMY-2
CUMC223	E. coli	6 (R)	13 (+)	13 (R)	13 (−)	9.2	CMY-2	OP2	CMY-2
CUMC224	E. coli	7 (R)	13 (+)	15 (I)	12 (−)	9.2, 5.4	CMY-2	NH4	CMY-2, TEM-1-like
CUMC225g	E. coli	6 (R)	11 (+)	11 (R)	11 (−)	9.2	CMY-2	NH2	CMY-2
CUMC245	E. coli	16 (I)	29 (+)	6 (R)	30 (+)	9.3, 7.4, 6.0	CTX-M-15, OXA-30	OP3	CTX-M-15, OXA-30, pI 6.0
CUMC174f	E. coli	27 (S)	30 (−)	16 (I)	30 (+)	8.0, 5.4	CTX-M-14, TEM-1-like	NH1	CTX-M-14, TEM-1-like
CUMC176h	E. coli	30 (S)	33 (−)	17 (I)	31 (+)	8.0, 5.4	CTX-M-14, TEM-1-like	OP1	CTX-M-14, TEM-1-like
CUMC180	E. coli	29 (S)	31 (−)	17 (I)	30 (+)	8.0, 5.4	CTX-M-14, TEM-1-like	NH1	CTX-M-14, TEM-1-like
CUMC192f	E. coli	28 (S)	30 (−)	15 (I)	30 (+)	8.0, 5.4	CTX-M-14-like, TEM-1-like	NH1	CTX-M-14-like, TEM-1-like
CUMC203f	E. coli	28 (S)	29 (−)	17 (I)	32 (+)	8.0, 5.4	CTX-M-14-like, TEM-1-like	NH1	CTX-M-14-like, TEM-1-like
CUMC204f	E. coli	27 (S)	30 (−)	15 (I)	30 (+)	8.0, 5.4	CTX-M-14-like, TEM-1-like	NH1	CTX-M-14-like, TEM-1-like
CUMC233	E. coli	31 (S)	32 (−)	18 (I)	32 (+)	8.0, 5.4	CTX-M-14-like, TEM-1-like	NH1	CTX-M-14-like, TEM-1-like
CUMC241h	E. coli	28 (S)	32 (−)	13 (R)	30 (+)	8.0, 5.4	CTX-M-14, TEM-1-like	NH6	CTX-M-14, TEM-1-like
CUMC244h	E. coli	26 (S)	29 (−)	14 (R)	29 (+)	8.0, 5.4	CTX-M-14, TEM-1-like	OP1	CTX-M-14, TEM-1-like
CUMC248f	E. coli	25 (S)	29 (−)	12 (R)	28 (+)	8.0, 5.4	CTX-M-14, TEM-1-like	NH1	CTX-M-14, TEM-1-like
CUMCK9	K. pneumoniae	6 (R)	26 (+)	18 (I)	32 (+)	8.2, 7.6	SHV-like	NH2	SHV-12-like, SHV-1-like

^aSusceptibilities were determined by disk diffusion. CAZ, ceftazidime; CAZ-CA, ceftazidime plus clavulanate (10 μg); CTX, cefotaxime; CTX-CA, cefotaxime plus clavulanate (10 μg). Numbers for disk diffusion represent zone sizes in millimeters. S, I, and R, susceptible, intermediate, and resistant, respectively, using CLSI criteria. The criterion for a positive ESBL confirmatory disk test is an increase in zone size of ≥5 mm for either CTX or CAZ in the presence of CA compared with either drug alone. +, ≥5-mm increase in zone size; −, <5-mm increase in zone size.

^bIEF point(s) of the beta-lactamase(s) identified by IEF.

^cMolecular analysis was performed by PCR and, in some cases, sequence analysis of full-length amplicons. “like” indicates PCR data only.

^dLocation of the patient from which the isolate was collected. Numbers represent the site in case of multiple locations. NH, nursing home; OP, outpatient clinic, IP, hospitalized patient.

^eAmpC, ESBL, or other β-lactamases produced by the isolate or, in the case of nonsequenced genes, the most likely types of enzymes produced by the isolate based on IEF and PCR results. The β-lactamase with a pI of 6.0 could not be identified further.

^fIsolates were obtained from the same patient.

^gIsolates were obtained from the same patient.

^hIsolates were obtained from the same patient.

Twelve E. coli isolates (16% of NS-FOX/CPD strains and 0.8% of the total) and two Klebsiella pneumoniae isolates (14% of NS-FOX/CPD strains and 0.9% of the total) produced ESBLs, while nine E. coli isolates (12% of NS-FOX/CPD strains and 0.6% of the total) and one K. pneumoniae isolate (CUMCK2) (7% of NS-FOX/CPD strains and 0.5% of the total) produced a CMY-2 imported AmpC (Table 1). Seventy percent (7/10) of the AmpC producers came from nursing homes, while 20% (2/10) and 10% (1/10) were collected from outpatient clinics and the hospital, respectively. Seventy-nine percent (11/14) of ESBL producers were collected from nursing homes, whereas 21% (3/14) were collected from outpatient clinics. Two isolates, a K. pneumoniae isolate (CUMCK2) and an E. coli isolate (CUMC215), from two patients in one nursing home produced both a CMY-2 and a CTX-M-14 β-lactamase. In addition, an E. coli isolate, CUMC245, from an outpatient clinic produced a CTX-M-15 and an OXA-30 β-lactamase (Table 1).

Although nosocomial infections are the major emphasis of surveillance studies regarding gram-negative pathogens producing ESBLs and AmpC β-lactamases, there is an increasing incidence of these organisms in community settings and LTCFs. The lack of surveillance for these organisms from LTCF and community patients can lead to an increased incidence of nosocomial infections when infected or colonized patients are admitted to the hospital (22).

To our knowledge, this study is the most recent U.S. study identifying the types of ESBLs and imported AmpC β-lactamases from urine isolates from both LTCF and clinic patients. In one LTCF, two different patients harbored two species that coproduced an imported CMY-2 AmpC and a CTX-M-14 β-lactamase. In the patient who harbored the K. pneumoniae isolate producing both β-lactamases, an E. coli strain (CUMC174) was isolated from the same urine sample and produced the CTX-M-14 β-lactamase but not the AmpC. It is interesting to note that this patient was treated with mandelamine 3 days after the urine sample was analyzed and subsequent urine samples from that patient over a 6-month period yielded E. coli producing CTX-M-14, but K. pneumoniae was not subsequently isolated (Table 1).

The majority of ESBLs in this study were CTX-M β-lactamases, while all the imported AmpC enzymes were CMY-2. These data are consistent with a recent study from Texas in which most of the ESBLs were CTX-M β-lactamases, with CTX-M-15 predominating (13). In the present study, however, the dominant CTX-M β-lactamase was CTX-M-14, with only 1 out of 12 isolates coproducing CTX-M-15 and OXA-30. This is the first documented account in the United States of an OXA-30 β-lactamase being coproduced with a CTX-M ESBL in an E. coli isolate collected from a clinic patient. Although OXA-30 is not characterized as an ESBL, the enzyme is capable of hydrolyzing cefepime, and its production in a community-acquired isolate is a concern (11).

The nonhospital patient study isolates came from five nursing homes and four outpatient clinics. Of urine isolates producing ESBLs or AmpC β-lactamases, 23% (5/22) came from outpatient clinics. Seventy percent (7/10) of the AmpC producers came from nursing home patients, while 20% (2/10) were collected from clinic patients. Seventy-nine percent (11/14) of ESBL producers were collected from nursing home patients, whereas 21% (3/14) were collected from clinic patients. Although the original sources of these isolates are unknown, these data suggest that ESBL and imported AmpC β-lactamase producers are not uncommon in nonhospitalized patients such as those from LTCFs and outpatient clinics.

Current susceptibility assays are not sufficient for accurate surveillance of ESBL- and AmpC-producing pathogens. Seven of the AmpC-producing isolates in this study (CUMC243, -181, -190, -214, -223, -224, and -225) were falsely identified as ESBL producers by disk diffusion but not by microbroth testing. There are no criteria described by the CLSI for the detection of AmpC-producing pathogens. In this study, additional phenotypic testing was required in conjunction with molecular testing to identify the types of β-lactamases produced by these strains. Laboratories actively testing for ESBL- and AmpC-producing organisms need to be aware that some of these organisms may test falsely positive for ESBLs using CLSI methodology. Further studies are required to determine the extent and impact of this problem.

Infections caused by antibiotic-resistant organisms continue to increase in LTCFs (10). This trend, combined with the potential for increased incidence of community-acquired ESBL/AmpC-producing bacteria, may lead to an increased risk for the spread of these resistant organisms to hospitalized patients. There is a need to broaden surveillance studies to include the community that makes up the patient population of area hospitals. However, the methodologies used for detection of ESBL/AmpC-producing organisms also need to be reviewed and improved. Acknowledging that community-based patients may serve as reservoirs for resistant organisms may be important when determining what infection control precautions are implemented when patients such as these are admitted to the hospital.