|Home | About | Journals | Submit | Contact Us | Français|
Seven carbapenem-resistant NDM-1-positive Klebsiella pneumoniae isolates were recovered from patients hospitalized between 2007 and 2009 in different wards at a referral and tertiary care center in Nairobi. Most of the isolates were obtained from urine. All isolates carried the blaNDM-1 carbapenemase gene previously reported from India, Pakistan, and the United Kingdom. These isolates were clonally related and expressed many other resistance determinants, including β-lactamases CTX-M-15, OXA-1, OXA-9, CMY-6, and aminoglycoside resistance methylase RmtC. This work corresponds to the first report of NDM-1 producers in Africa.
Carbapenem-hydrolyzing β-lactamases belonging to Ambler classes A, B, and D have been reported worldwide among Enterobacteriaceae (10, 14, 18). The most clinically significant are KPC-type (Ambler class A), IMP and VIM types (class B), and OXA-48 (class D), mostly identified in Klebsiella pneumoniae as a source of nosocomial outbreaks (14). Most of those isolates are multidrug resistant.
Recently, a novel metallo-β-lactamase (MBL) named NDM-1 (New Delhi metallo-β-lactamase) has been identified from K. pneumoniae (strain 05-506) and Escherichia coli isolates recovered in Sweden from a patient previously hospitalized in India (19). An extensive survey performed in the United Kingdom, India, and Pakistan identified NDM-1-producing K. pneumoniae, E. coli, Citrobacter freundii, Morganella morganii, Providencia spp., and Enterobacter cloacae isolates (9). The presence of NDM-1 producers in hospitalized patients in the United Kingdom was related in many cases to previous hospitalization in the Indian subcontinent (9), and similar observations have been made in France (12, 15) and the Sultanate of Oman (11). Spread of this novel carbapenemase gene is considered a serious threat since the reservoir of NDM-1 producers is at least in part related to the Indian subcontinent, which is inhabited by the second-largest population in the world and where NDM-1 producers are reported also in community-acquired infections (9, 12).
Our study was initiated using an archival collection of bacterial isolates from the Aga Khan University Hospital, Nairobi, Kenya, for the purpose of surveillance of antibacterial resistance mechanisms and infection control audits. Seven multidrug-resistant K. pneumoniae strains isolated over a 3-year period (2007 to 2009) were selected, all showing an identical pattern of resistance. Those isolates had been recovered mostly from urine, from patients who were all receiving antibiotic treatments when the samples had been recovered (Table (Table1).1). However, those treatments were not targeted toward the carbapenem-resistant K. pneumoniae isolates that were systematically considered non-strictly invasive (Table (Table11).
The MICs were determined by Etest (AB bioMérieux, Solna, Sweden) on Mueller-Hinton agar plates and by liquid microdilution assays at 37°C, and results of susceptibility testing were interpreted according to the CLSI guidelines (6). The isolates were resistant to all β-lactams, including carbapenems with MIC values of >32 μg/ml for imipenem, meropenem, ertapenem, and doripenem. In addition, they were resistant to aminoglycosides, fluoroquinolones, chloramphenicol, sulfonamides, fosfomycin, and nitrofurantoin, the MIC of rifampin was >32 μg/ml, and the MICs of tigecycline and colistin measured by Etest were 0.5 and 0.5 μg/ml, respectively.
Results from MBL detection performed by using Etest MBL strips (AB bioMérieux) were positive. Thus, PCR assays were carried out with a series of primers designed for the detection of several class B β-lactamase genes, blaIMP, blaVIM (16), and blaNDM-1 (primer NDM-Fm, 5′-GGTTTGGCGATCTGGTTTTC-3′; and primer NDM-Rm, 5′-CGGAATGGCTCATCACGATC-3′). The isolates were positive for blaNDM-1, and sequencing of the PCR products revealed 100% identity with the published sequence of the blaNDM-1 gene (19). Noteworthy, all the patients from whom the NDM-1-producing isolates had been recovered were Kenyans living in Kenya, but the history of their travel or contact with Indian or British populations could not be recovered. All these isolates were identified from hospitalized patients, and pulsed-field gel electrophoresis (PFGE) was performed as described previously (3) to evaluate any clonal relationship. PFGE analysis showed an identical pattern for all isolates (Fig. (Fig.1).1). The hospitalization wards from which the infected patients were originating were located in distinct locations inside the hospital, and no source or index cases could be identified. Very interestingly, the PFGE pattern of those Kenyan isolates was very similar to that of the first reported NDM-1-producing K. pneumoniae strain 05-506 identified in Sweden from a patient hospitalized in India in 2008 (Fig. (Fig.1)1) (9). It was also very similar to that of K. pneumoniae strain 601, recently identified in the Sultanate of Oman (Fig. (Fig.1),1), for which a link with India was evidenced (11). This result indicates a clonal link between those isolates identified in Nairobi and this Indian strain. That clonal link was further confirmed by performing multilocus sequence typing as described previously (7) which showed that the K. pneumoniae isolates of Kenya belonged to the ST14 sequence type, as reported for K. pneumoniae 05-506 (19).
Since high-level resistance to cephalosporins and monobactam aztreonam (a substrate spared by NDM-1) was observed for all isolates, detection of extended-spectrum β-lactamase (ESBL) and AmpC productions was carried out by PCR analysis using specific primers to detect broad-spectrum β-lactamase genes (4), followed by sequencing. All isolates harbored the blaSHV-28 and blaCTX-M-15 ESBL genes, together with the blaCMY-6 AmpC gene and the blaOXA-1 and blaOXA-9 class D β-lactamase genes. Yong et al. (19) reported that the NDM-1-positive K. pneumoniae strain 05-506 was blaCMY-4 positive. Screening of 16S rRNA genes encoding methylase, performed by using a multiplex PCR approach as described previously (1), identified the rmtC methylase gene conferring high-level resistance to all aminoglycosides in the Kenyan isolates.
Transferability of the blaNDM-1 gene was studied by conjugation experiments as described previously (17), with a selection based on ceftazidime (30 μg/ml) and azide (100 μg/ml), using K. pneumoniae Kp7 as the donor strain and E. coli J53 (resistant to azide) as the recipient strain. One of the obtained transconjugants, E. coli(pKp7-A) expressing NDM-1, had decreased susceptibility to carbapenems (MIC values of imipenem, meropenem, ertapenem, and doripenem were 4, 1.5, 1.5, and 1.5 μg/ml, respectively [MICs for these same molecules were 0.12 μg/ml for the E. coli J53 recipient strain]). It was resistant to all aminoglycosides and to sulfonamides, and PCR analysis confirmed that plasmid pKp7 coharbored the rmtC gene. This plasmid, analyzed by using the Kieser technique (8), was 120 kb in size and belonged to the IncA/C2 incompatibility group as demonstrated by the PCR-based replicon typing method (2), whereas the blaNDM-1-positive plasmid in K. pneumoniae 05-506 was untypeable (9), and that of K. pneumoniae 601 from the Sultanate of Oman was reported to belong to the IncL/M group (11). Kumarasamy et al. (9) reported blaNDM-1-positive plasmids from the Indian isolates that were either untypeable or belonged to the IncA/C or IncFI/FII plasmid incompatibility group. Another type of transconjugant was obtained, E. coli(pKp7-B), expressing an ESBL phenotype. It harbored the blaCTX-M-15 and blaTEM-1 genes, had a 75-kb plasmid, belonged to the InF group, and did not carry any other non-β-lactam resistance markers.
PCR mapping was performed to identify the genetic sequences surrounding the blaNDM-1 gene in K. pneumoniae Kp7 for comparison with the sequences surrounding the blaNDM-1 gene in K. pneumoniae 05-506 (19). PCR experiments based on primers located in those surrounding sequences produced negative results, indicating that the structures identified in K. pneumoniae 05-506 were not present on plasmid pKp7.
Our study further underlines the spread of the blaNDM-1 gene worldwide, as exemplified by the newly reported NDM-1-producing E. coli isolate identified in Australia (13), together with the report of NDM-1-producing E. coli, K. pneumoniae, and Enterobacter cloacae in the United States (5). This work is the first to report the dissemination of that gene in Africa which actually corresponds to the very first known identification of an NDM-1 producer. It has been traced to 2007, prior to the first known identification of an NDM-1 producer in 2008 in Sweden (19). Histories of contacts or the origin of the Kenyan patients to the Indian subcontinent could not be evidenced. However, taking into account the size of the Indian diaspora in Kenya (100,000 people), it remains possible that introduction of NDM-1-producing K. pneumoniae in the Kenyan population could be linked to the Indian subcontinent. At the same time it may be speculated that the large international community traveling in and out of Kenya and the air traffic between Europe, the United Kingdom, and East Africa could have easily facilitated such an event. The clonal relationship between the NDM-1-positive K. pneumoniae isolate 05-506 from Sweden, the K. pneumoniae isolate 601 from the Sultanate of Oman, and the Kenyan isolates presents a very strong argument that this link might exist. Interestingly, K. pneumoniae isolates belonging to the same clonal lineage were found to harbor different blaNDM-1-positive plasmids. This is in accordance with the observations made by Kumarasamy et al. (9), who showed an important diversity of plasmid size and scaffolds among NDM-1-positive isolates recovered from the United Kingdom, India, and Pakistan.
Finally, this report described the first occurrence of NDM-1 producers in Africa. However, this may not be considered a nosocomial outbreak since these isolates were recovered over a long period of time and from patients who did not have any obvious epidemiological connection. The multidrug resistance pattern of the K. pneumoniae isolates identified here mirrors that of NDM-1 producers previously identified in India, Pakistan, and the United Kingdom and is a source of deep concern. As extensively reported for outbreaks of ESBL producers, K. pneumoniae is a bacterial species that is prone to being an excellent vector for hospital dissemination of resistance genes. We fear that gene exchanges similar to those observed extensively for ESBL genes may occur for the blaNDM-1 gene, i.e., gene circulation in community-acquired E. coli which is then transferred to hospital-based K. pneumoniae.
This work was funded by a grant from the INSERM (U914), Paris, France, by the Ministère de l'Education Nationale et de la Recherche (UPRES-EA3539), by Université Paris XI, Paris, France, and by a grant from the European Community (TEMPOtest-QC, HEALTH-2009-241742).
We thank W. Liu for technical assistance. We thank T. R. Walsh for the gift of the K. pneumoniae 05-506 control strain.
Published ahead of print on 29 November 2010.