New Delhi Metallo-β-Lactamase (NDM-1)-Producing Klebsiella pneumoniae: Case Report and Laboratory Detection Strategies

Abstract

The spread of antimicrobial resistance among Enterobacteriaceae is a significant clinical threat. We report the first case of an Enterobacteriaceae strain harboring the NDM-1 metallo-β-lactamase in a pediatric patient in the United States. We describe strategies for the detection of this novel resistance mechanism encountered in an isolate of Klebsiella pneumoniae.

CASE REPORT

A 13-month-old boy in severe respiratory distress was brought to the emergency department (ED) by his parents following a 3-day history of worsening cough and audible wheezing. At presentation, the patient was tachycardic and lethargic and had a low-grade fever of 38.1°C. The patient's past medical history was notable for episodes of reactive airway disease, managed with nebulized albuterol and short courses of corticosteroids, congenital skeletal abnormalities, multiple upper respiratory tract infections, hydrocele, hypotonia, and developmental delay. Also of note, the patient had returned from a 3-month visit to Pakistan with his family 4 months prior to presentation in the ED. While in Pakistan, the patient was hospitalized following an exacerbation of his lung disease and was intubated and mechanically ventilated for several days.

The patient was intubated in the ED, started on empirical ceftriaxone (550 mg intravenously [i.v.] daily) for pneumonia, and admitted to the pediatric intensive care unit. On admission, a chest X-ray revealed right upper and lower lobe opacities, and azithromycin treatment (60 mg i.v. daily) was started. A urine specimen was sent to the microbiology laboratory for culture, and a Foley catheter was placed. Urinalysis revealed no pyuria, but nitrites were detected and less than 10,000 CFU of Klebsiella pneumoniae (strain KP1) and Proteus species grew in routine culture; these organisms were also isolated in similar quantities from a subsequent urine specimen obtained later that day. A nasal wash specimen was obtained for bacterial culture and respiratory virus detection by the Resplex II respiratory virus panel, version 2.0 (Qiagen, Valencia, CA). The respiratory virus panel detected the presence of an enterovirus. Following overnight incubation at 37°C in air supplemented with 5% CO₂, bacterial culture grew normal oral flora and many K. pneumoniae organisms (strain KP2). Following identification of KP2, empirical antimicrobial coverage was switched from ceftriaxone to meropenem (MEM; 440 mg i.v. every 8 hours [q8h]) on hospital day 4. Also on hospital day 4, the patient's tracheal secretions increased in amount, requiring more-frequent suctioning to maintain a patent airway. A small right pleural effusion was detected by a computed tomography scan on hospital day 4, which subsequently resolved without drainage. On hospital day 5, respiratory secretions obtained by tracheal suction were submitted for bacterial culture, which also grew a K. pneumoniae isolate (KP3). No additional specimens were sent to the laboratory for culture.

Antimicrobial susceptibility testing was performed on all three K. pneumoniae isolates by use of the Clinical and Laboratory Standards Institute (CLSI) reference broth microdilution method (2) on in-house-prepared panels. KP1 tested positive for extended-spectrum β-lactamases (ESBL) with the use of the CLSI disk diffusion method (5). KP1 was susceptible to ertapenem (ETP), imipenem (IMP), and MEM, with MICs at ≤0.25, ≤0.25, and ≤0.25 μg/ml, respectively. In contrast, the MICs for ETP, IMP, and MEM for KP2 were considerably higher at 8, 4, and 8 μg/ml, respectively, and those for KP3 were 8, 8, and 16 μg/ml, respectively (Table 1). All three isolates were resistant to aminoglycosides, ciprofloxacin, and trimethoprim-sulfamethoxazole; however, KP2 and KP3 were resistant to all β-lactams tested, which was not the case for KP1 (Table 1). Colistin MICs were determined for KP2 and KP3 by broth macrodilution (2) and by Etest; KP2 had an MIC of 0.25 μg/ml, whereas KP3 had an MIC of 32 μg/ml. On hospital day 5, upon receipt of antimicrobial susceptibility testing (AST) results for KP2, azithromycin and MEM treatment was discontinued, and the patient was started on colistin (15 mg i.v. twice a day [BID]), which was continued for a 10-day course. The endotracheal tube was removed on hospital day 8, and the patient was discharged home on hospital day 12.

Table 1.

Broth microdilution MICs for K. pneumoniae NDM-1 strains isolated from urine (KP1), nasal wash (KP2), and sputum (KP3) samples^a

Antibiotic	KP1		KP2		KP3
	MIC (μg/ml)	Interpretationb	MIC (μg/ml)	Interpretation	MIC (μg/ml)	Interpretation
Amikacin	>32	R	>32	R	>32	R
Aztreonam	>32d	R	>32	R (R)	>32	R
Cefepime	DD	S	>32	R	>32	R
Ceftazidime	32		>32	R	>32	R
Ceftriaxone	>32	R	>32	R	>32	R
Ciprofloxacin	>2	R	>2	R	>2	R
Colistin	ND		0.252	NG	32c	NG
Ertapenem	≤0.25	S (S)	8	R (R)	8	R (R)
Gentamicin	>10	R	>10	R	>10	R
Imipenem	≤0.25	S (S)	4	R (S)	8	R (I)
Meropenem	≤0.25	S (S)	8	R (I)	16	R (R)
Minocycline	ND		4	S	≤4	S
Piperacillin-tazobactam	≤8	S	>128	R	>128	R
Tigecycline	ND		1	S	≤0.5	S
Tobramycin	>10	R	>10	R	>10	R
Trimethoprim-sulfamethoxazole	>4/80	R	>4/80	R	>4/80	R

^aND, not determined; DD, disk diffusion; NG, no interpretive criteria available.

^bResults obtained using CLSI 2010 M100-S20-U, June 2010 revised MIC interpretative criteria for strains susceptible (S), intermediate (I), and resistant (R) to ertapenem, IMP, and meropenem, which are as follows: for ertapenem, ≤0.25, 0.5-1, and ≥2, respectively; for IMP, ≤1, 2, and ≥4, respectively; and for meropenem, ≤1, 2, and ≥4, respectively (4). In parentheses are results obtained using CLSI M100-S20, January 2010 MIC (μg/ml) interpretative criteria for strains susceptible, intermediate, and resistant to ertapenem, imipenem, and meropenem, which are as follows: for ertapenem, ≤2, 4, and ≥8, respectively; for imipenem, ≤4, 8, and ≥16, respectively; and for meropenem, ≤4, 8, and ≥16, respectively (5).

^cMacrobroth tube dilution method.

^dVitek GNS-137 card.

A modified Hodge test (MHT) was performed on KP2 and KP3, using both MEM and ETP (5). KP2 was MHT positive in both tests, whereas KP3 was MHT negative (Fig. 1A). Given the patient's history of travel to Pakistan, the isolates were evaluated phenotypically for the presence of a metallo-β-lactamase (MBL), using the metal chelating agent EDTA. Identification of MBL activity was performed by three methods: a carbapenem-EDTA combined disk method (11), a Tris-EDTA (EDT) double-disk synergy method (11), and an MBL Etest (bioMérieux, Durham, NC). For all tests, Stenotrophomonas maltophilia (strain ATCC 51331), which is known to intrinsically express a chromosomal metallo-β-lactamase (9), was used as a positive control. Briefly, the combined disk test was performed by inoculating the isolates onto Mueller-Hinton agar (MHA) per CLSI guidelines for a standard disk diffusion test (3). Two imipenem (IMP; 10 μg) disks were applied to the inoculated plates, and 10 μl of a sterile 0.5 M EDTA (pH 8.0) solution was applied to one disk. A sterile 6-mm filter paper disk to which 10 μl of 0.5 M EDTA was applied was used to determine if EDTA alone might inhibit the growth of the test isolates. The plates were incubated at 35°C under ambient air for 18 h. The zones of inhibition around the IMP and IMP-EDTA disks were measured, and zone increases of ≥7 mm for both KP2 and KP3 in the presence of EDTA were noted and interpreted as indicative of an MBL phenotype on the basis of criteria described previously (6). Both KP2 and KP3 grew up to the EDTA disk. The EDTA combined disk test was repeated using both MEM (10 μg) and ETP (10 μg) disks, and KP2 and KP3 yielded positive results for all combinations (Table 2). Figure 1B illustrates the MBL activity of KP2 as determined using the carbapenem-EDTA combined disk method. The EDT double-disk synergy method also yielded positive results for both KP2 and KP3. In this assay, a BBL Sensi-disk EDT disk (BD, Sparks, MD) was placed adjacent to an IMP disk; an enhanced zone of inhibition toward the EDT disk was read as a positive result (Fig. 1C). MBL Etest strips (bioMérieux, Durham, NC) containing IMP alone and IMP-EDTA were tested according to the manufacturer's recommendations. The results for both KP2 and KP3 were difficult to interpret due to large zones of inhibition surrounding the entire Etest strips (Fig. 1D). However, when compared to a negative control (Fig. 1E), a small distortion in the ellipse of growth inhibition in the middle of the Etest strip was noted (Fig. 1D, arrowheads), which is indicative of a positive test result. The genetic relatedness of the three K. pneumoniae isolates from the patient was determined by repetitive sequence-based PCR (rep-PCR) using a semiautomated system (DiversiLab, bioMérieux, Durham, NC) and a DiversiLab Klebsiella kit (bioMérieux, Durham, NC). KP2 and KP3 were indistinguishable, with band patterns 97% identical, whereas KP1 was unrelated (<90% similarity) to KP2 and KP3 (data not shown).

Fig. 1.

Phenotypic detection of carbapenemases in a K. pneumoniae NDM-1 strain isolated from a nasal wash specimen (KP2). (A) Modified Hodge test, using ETP, for detection of carbapenemase, performed by the CLSI reference method. (B) Carbapenem-EDTA combined disk test for MBL. The K. pneumoniae isolate was inoculated to the surface of a Mueller-Hinton agar plate. Disks containing 10 μg/ml ETP, IMP, and MEM alone or ETP, IMP, and MEM plus 10 μl of 0.5 M EDTA (pH 8.0) were placed on the agar. EDTA was also added to a blank disk as a control. (C) EDTA-IMP disk synergy test for MBL. (D) Equivocal MBL Etest result demonstrating a large zone of inhibition. Arrowheads indicate the deformation of the ellipse (i.e., phantom zone). (E) Negative (control) MBL Etest results demonstrating a characteristic sharp-edged ellipse and a large zone of inhibition.

Table 2.

Results of the carbapenem-EDTA combined disk diffusion test for MBL^a

Isolate	Zone size (mm) for:
	ETP	ETP-EDTA	MEM	MEM-EDTA	IMP	IMP-EDTA
KP2	10	21	11	25	10	24
KP3	14	22	15	27	17	24

^aETP, ertapenem; ETP-EDTA, ertapenem-EDTA; MEM, meropenem; MEM-EDTA, meropenem-EDTA; IMP, imipenem; IMI-EDTA, imipenem-EDTA.

KP2 and KP3 were sent via the Los Angeles County Public Health Department to the Centers for Disease Control and Prevention, where the presence of an NDM-1 gene in KP2 and KP3 was confirmed by PCR (8). As these isolates were the first seen in the United States in a patient with history of travel to Pakistan, the bla_NDM-1 gene was sequenced, which yielded results identical to the sequence obtained from the other three NDM-1-producing Enterobacteriaceae isolated in the United States (8) and indistinguishable from the published bla_NDM-1 sequence (NCBI accession number FN396877). Pulsed-field gel electrophoresis of the isolates performed by the CDC confirmed that KP2 and KP3 were indistinguishable and revealed a pattern unrelated to that for the other NDM-1 producing K. pneumoniae strains isolated in the United States and unlike that for any common K. pneumoniae strain seen in the United States (data not shown). MBL activity was confirmed at the CDC using the MBL Etest, whose results were interpreted as positive for both KP2 and KP3. The results for repeated ETP MHT assays were interpreted at the CDC as positive for both KP2 and KP3; the results for repeated MEM MHT assays were negative for both isolates.

NDM-1 is a novel MBL (12) encoded by the bla_NDM-1 gene, which confers resistance to all β-lactams, with the exception of aztreonam (12). Many strains that harbor bla_NDM-1 are also aztreonam resistant, presumably by a different resistance mechanism (7). Since the original report, bla_NDM-1-harboring isolates have widely been identified across the Indian subcontinent and in the United Kingdom. Half (46%) of cases reported to occur in the United Kingdom are associated with travel to India or Pakistan (7). More recently, NDM-1 cases have been identified across the globe, including in Canada, Europe, Japan, Africa, and Australia (10), and all of these cases are associated with travel to or medical care in India or Pakistan. In the United States, three NDM-1-producing isolates were previously confirmed, including an Escherichia coli isolate, a K. pneumoniae isolate, and an Enterobacter cloacae isolate. All three isolates were from patients with recent exposure to medical care in India and were isolated from urine specimens (1). This report describes the fourth case of an NDM-1-producing Enterobacteriaceae in the United States and specifically outlines laboratory strategies for the detection of an isolate with this important new mechanism of carbapenem resistance.

MHT is the only CLSI-recommended phenotypic assay for detection of carbapenemase production in Enterobacteriaceae. The MHT result for KP2 was not as strongly positive as it was for the control strain (Fig. 1A), a result that some laboratories may interpret as negative, and the MHT result for KP3 was negative. Similar unreliable performance of the MHT for detection of the presence of carbapenemase has been noted for one of the three other US NDM-1 isolates (P. Schrekenberger, unpublished data, and reference 1). Phenotypic assays other than the MHT have been described in the literature for the detection of MBL (6). Three of those assays performed by our laboratory on KP2 and KP3 yielded results consistent with MBL expression by these isolates. In our hands, we felt the IMP-EDTA combined disk method was the most simple to perform and to interpret (6); however, it is important to note that no data exist to suggest that future NDM-1 isolates or other MBL types will be detected by these methods.

In June 2010, the CLSI revised the Enterobacteriaceae carbapenem breakpoints based on contemporary microbiological, pharmacologic, and clinical outcome data (4). These changes encompassed significant decreases in the MIC breakpoints for susceptibility to the carbapenems and obviate the performance of the MHT for guiding patient treatment (4). Phenotypic and/or molecular assays may be employed when identification of the carbapenem resistance mechanism is of interest; however, limitations of phenotypic assays must be understood.

Like many NDM-1 producing isolates, our patient's KP2 and KP3 isolates were resistant to all conventional antimicrobial agents, with the exceptions of tigecycline and minocycline, which we were averse to using in a pediatric patient. While KP2 had a low MIC for colistin (0.25 μg/ml), KP3, which was isolated on the same day as colistin therapy was started, had a colistin MIC of 32 μg/ml. Our patient recovered from his pneumonia while on colistin monotherapy, and so the significance of isolating a colistin-resistant organism in this case is questionable. Furthermore, as no follow-up cultures were performed on our patient following the resolution of pneumonia, it is possible the patient remains colonized with the NDM-1 K. pneumoniae.

In summary, we report herein the first case of pneumonia caused by K. pneumoniae harboring the bla_NDM-1 gene in the United States. The significant risk factor for this infection was prior hospitalization in Pakistan. Based on our experience, we feel that the most cautious and expedient method for laboratories to identify carbapenem resistance is by implementing the new CLSI carbapenem breakpoints (4), if possible. Additionally, laboratories may consider sending carbapenem-resistant Enterobacteriaceae from patients with travel history including receipt of medical care to the CDC via their local public health department for further evaluation. Communication between the laboratory and clinicians regarding risk factors, such as travel history, for patients with carbapenem-resistant Enterobacteriaceae is essential in identifying NDM-1 and other new resistance phenotypes.