ORIGINAL ARTICLE 10.1111/j.1469-0691.2008.02030.x Metallo-b-lactamase-producing Pseudomonas aeruginosa isolated from a large tertiary centre in Kenya J. D. D. Pitout 1,2,3, G. Revathi 4, B. L Chow 1, B. Kabera 4, S. Kariuki 5, P. Nordmann 6 and L. Poirel 6 1 Division of Microbiology, Calgary Laboratory Services, 2 Department of Pathology & Laboratory Medicine, 3 Microbiology and Infectious Diseases, University of Calgary, Calgary, Alberta, Canada, 4 Division of Microbiology, Department of Pathology, Aga Khan University Hospital, 5 Centre for Microbiology Research, Kenya Medical Research Institute, Nairobi, Kenya and 6 Service de Bactériologie-Virologie, INSERM U914 Emerging Resistance to Antibiotics, Hôpital de Bicêtre, Assistance Publique Hôpitaux de Paris, Faculté de Médecine et Université Paris-Sud, K.-Bicêtre, France ABSTRACT This study was designed to characterize the b-lactamase content of carbapenem-resistant Pseudomonas aeruginosa isolates recovered during 2006 and 2007 in a large tertiary-care centre in Nairobi, Kenya. Molecular characterization was done using PCR and sequencing, and typing was performed using pulsed-field gel electrophoresis (PFGE). In total, 416 P. aeruginosa isolates were obtained during that period, of which 57 (13.7%) were resistant to carbapenems. All carbapenem-resistant isolates tested positive for metallo-b-lactamase (MBL) production. All MBL isolates produced VIM-2 with two types of integron structures. PFGE identified three clonally related groups of VIM-2-producing P. aeruginosa, including a pan-resistant clone that was responsible for nosocomial outbreaks during 2006 and 2007 in the intensive-care unit. These findings suggest that continuous molecular surveillance needs to be performed to monitor the spread within the hospital of this pan-resistant strain. This study is the first report of VIM-2-producing P. aeruginosa from the African continent. Keywords Kenya, Metallo-b-lactamases, Pseudomonas aeruginosa Original Submission: 3 April 2008; Revised Submission: 29 April 2008; Accepted: 5 May 2008 Edited by A. Sundsfjord Clin Microbiol Infect 2008; 14: 755 759 INTRODUCTION Pseudomonas aeruginosa isolates producing metallob-lactamases (MBLs) were first reported from Japan in 1991, and since then have been described from various parts of the world, including Asia, Europe, Australia, South America and North America. [1] MBLs are binuclear zinc-dependent enzymes and are susceptible to inhibition by metal chelators such as EDTA. These enzymes have the ability to hydrolyze a wide variety of b-lactam agents such as penicillins, cephalosporins and carbapenems, but not the monobactams. The genes responsible for the production of MBLs are typically part of an integron structure Corresponding author and reprint requests: J. D. D. Pitout, Calgary Laboratory Services, #9, 3535 Research Road NW, Calgary, Alberta, Canada T2L 2K8 E-mail: johann.pitout@cls.ab.ca and are encoded on large transferable plasmids [2]. Therefore, P. aeruginosa isolates producing MBLs are often resistant to different groups of antimicrobial agents, and this resistance can be transferred to various types of bacteria. Acquired MBLs can be divided into six categories according to their molecular structure, namely the IMP, VIM, GIM, SIM and SPM types [1] and, more recently, AIM [3]. MBL-producing P. aeruginosa isolates have been responsible for several nosocomial outbreaks in tertiary centres from different parts of the world, illustrating the need for proper infection control practices [4]. These isolates have also been responsible for serious infections such as septicaemia and pneumonia, and have been associated with failure of therapy with carbapenems [5]. The Division of Microbiology at the Aga Khan University Hospital in Nairobi, Kenya noted the Journal Compilation Ó 2008 European Society of Clinical Microbiology and Infectious Diseases
756 Clinical Microbiology and Infection, Volume 14 Number 8, August 2008 appearance of multiresistant P. aeruginosa in the intensive-care unit (ICU) during 2005. These isolates were resistant to the cephalosporins, carbapenems, aminoglycosides and fluoroquinolones. As MBL-producing Pseudomonas spp. have not previously been described from the African continent, a study was designed to determine whether MBLs were present among these isolates and to investigate the prevalence of these enzymes among clinical isolates of carbapenemintermediate or carbapenem-resistant (CR) P. aeruginosa from the Aga Khan University Hospital during 2006 2007. MATERIALS AND METHODS Aga Khan University Hospital Established in 1958, the Aga Khan University Hospital, Nairobi is a 254-bed long-term care facility that provides general medical services, specialist clinics and high-technology diagnostic services. The hospital serves as a referral centre for specialized medical care and diagnostic services from various hospitals and clinics within the region. Bacterial isolates Consecutive non-duplicate isolates of P. aeruginosa that were intermediate or resistant (non-susceptible) to either imipenem (MIC >8 mg L) or meropenem (MIC >8 mg L) (referred to as CR), collected at the Division of Microbiology, from the Aga Khan University Hospital during January 2006 to June 2007 were included in this study. Isolates were identified to the species level with Vitek (Vitek AMS; biomérieux Vitek Systems Inc., Hazelwood, MO, USA). Antimicrobial susceptibility testing MICs of the following drugs were initially determined by Vitek 2 (Vitek AMS; biomérieux Vitek Systems Inc.), and microdilution panels were performed on all MBL-positive isolates (Microscan Gram negative NMIC30; Dade Behring Canada Inc., Mississauga, ON, Canada); piperacillin, ceftazidime, cefepime, aztreonam, imipenem, meropenem, ciprofloxacin, gentamicin, amikacin and tobramycin. Colistin susceptibility was determined using the Etest (AB BioDisk, Solna, Sweden). The quality control strains used for this part of the study were Escherichia coli ATCC 25922 and P. aeruginosa ATCC 27853. Throughout this study, results were interpreted using CLSI criteria for broth dilution [6]. Identification of MBLs The presence of MBLs was evaluated in the clinical isolates of CR P. aeruginosa using both an EDTA screen test [7] and the MBL Etest (AB BioDisk) according to the manufacturer s instructions. DNA template preparation and duplex PCR amplification for the simultaneous detection of the bla IMP and bla VIM b-lactamase genes were carried out on a Thermal Cycler 9700 instrument (Applied Bio-systems, Foster City, CA, USA) with primers and conditions as previously described [7]. Sequencing of the VIM genes was performed using the class 1 integron primers 5CS (5 -GGC ATC CAA GCA GCA AG-3 ) and 3CS (5 -AAG CAG ACT TGA CCT GA-3 ) in combination with IMP and VIM primers, respectively. Two different amplicons per isolate were sequenced in an overlapping fashion to ensure that the entire MBL allele was obtained. Automated sequencing was performed on the PCR products with the ABI Prism 3100 Genetic Analyzer (Applied Biosystems, Norwalk, CT, USA) as previously described, using the EQUENCE NALYSIS software. [8] The 3CS extremity of the integron was specifically targeted using the VIM-2-specific forward primer [7], with the 402-TniC-B reverse primer (5 -ACGTGCCGGTAGAGCGTCTG-3 ) [9]. Pulsed-field gel electrophoresis (PFGE) PFGE was performed on selected MBL-producing isolates, using a protocol previously described [8]. Gel images in tiff format were exported to BIONUMERICS software version 3.5 (Applied Maths, Sint-Martens-Latem, Belgium) for analysis. Comparisons for P. aeruginosa were made using the bandbased Dice coefficient, which is a binary coefficient measuring similarity, based upon common and different bands. Dendrograms were generated using the unweighted pair group method using arithmetic averages (UPGMA) with 1.5% position tolerance. DNA relatedness was calculated, on the basis of the Dice coefficient, and isolates were considered to be genetically related if the Dice coefficient correlation was 80% or greater, which corresponds to the possibly related (four to six bands difference) criteria of Tenover et al.[10]. RESULTS Bacterial isolates During the 18-month study period, 416 P. aeruginosa isolates were obtained at Aga Khan University Hospital, of which 57 (13.7%) were resistant to the carbapenems: three (5%) were isolated from urine, four (7%) from blood, 17 (30%) from wounds (purulent), 30 (53%) from respiratory tract specimens, and the remaining three (5%) from various other specimens. The majority of CR isolates (33 57 (58%)) originated from the ICU. Antimicrobial susceptibility The 57 clinical isolates included in this study were highly resistant to imipenem and meropenem (MICs, >256 and >32 mg L respectively) and 30 (53%) were resistant to piperacillin and aztreonam, whereas 57 (100%) were resistant to ceftazidime, cefepime, tobramycin, gentamicin, amikacin and ciprofloxacin. No colistin resistance was detected, although 12 isolates had an MIC of
Pitout et al. P. aeruginosa producing MBLs from Kenya 757 2mg L (i.e. the CLSI breakpoint for colistin). Paterson recently suggested definitions for multidrug-resistant and pan-resistant P. aeruginosa [11]; all the isolates were multidrug-resistant and 53% were pan-resistant. Identification of MBLs and cassette arrays All of the CR P. aeruginosa isolates were MBLpositive according to both phenotypic methods, and PCR identified bla VIM genes in all of them. Among the 57 isolates, 15 isolates were randomly selected for additional molecular characterization. These included 11 isolates from the ICU, three from general wards, and one from an outpatient. PCR with primers 5CS and 3CS, combined with VIM primers, amplified various amplicons, ranging from 600 bp to 1.2 kb. Sequence analysis of the products obtained with the VIM primers revealed 100% identity with the sequence of the bla VIM-2 allele in all the isolates [12]. Although five of the 15 isolates possessed bla VIM-2 as a single cassette, the other ten bla VIM-2 -positive isolates carried two gene cassettes, consisting of bla VIM-2 preceded by an aaca4 gene cassette (encoding the aminoglycoside acetyltransferase AAC(6 )-Ib). PCR with the remaining 42 isolates showed that the majority (29 42) possessed bla VIM-2 as a single cassette, whereas the other 13 bla VIM-2 -positive isolates carried two gene cassettes, consisting of bla VIM-2 preceded by an aaca4 gene cassette. However, the 3CS extremity of the bla VIM-2 -containing integron was not amplified, probably suggesting a Tn402-like structure, as recently identified in a French bla VIM-2 -positive isolate at the origin of an outbreak [9]. PCR with primers specific for the TniC of Tn402 was negative as well. The primer combination consisted of the VIM-2-specific forward primer and the 402-TniC-B reverse primer (5 -ACGTGCCGGTA GAGCGTCTG-3 ). The positive control used during this part of the experiment was strain 29 [9]. A PFGE analysis was performed as described previously [13], followed by an I-CeuI digestion and a Southern blot hybridization analysis [12] for the determination of the precise genetic location of the bla VIM-2 genes in AFC-03, AFC-01 and AFC- 08. Chromosomal DNAs from VIM-2-positive P. aeruginosa isolate COL-1 were included as a positive control [12]. The membrane was successively hybridized with probes for bla VIM-2 and 16S and 23S rrna genes, as described previously, thus demonstrating the chromosomal location of bla VIM-2 (data not shown). Attempts to transfer the b-lactam resistance marker into E. coli DH10B and P. aeruginosa AFC-03 by electroporation failed. PFGE PFGE identified three closely related groups of VIM-2-producing P. aeruginosa among the 15 isolates, designated clusters MBLA (n = 6), MBLAR (n = 2) (i.e. related to A), and a separate cluster designated MBLB (n = 7) (Fig. 1). The MBLA, MBLAR and MBLB isolates formed separate clusters, within which there was >80% similarity of the PFGE profiles (Fig. 1). The MBLAR isolates exhibited profiles with >70% similarity to MBLA, which indicates that MBLA and MBLAR are related, although the significance is unknown. Among the 11 isolates from the ICU, two belonged to cluster MBLA, two to cluster MBLAR and seven to cluster MBLB; the remaining four from general wards and the outpatient isolate belonged to cluster MBLA. It is noteworthy that all the isolates belonging to cluster MBLB were panresistant (being sensitive only to colistin), whereas those from clusters MBLA and MBLAR were multiresistant, being sensitive to piperacillin, aztreonam and colistin. Analysis of bla VIM-2 showed that two of six strains from cluster MBLA, and one of two from clusters MBLAR and MBLB (n = 7), consisted of bla VIM-2 preceded by an aaca4 gene cassette, whereas the remaining strains from clusters MBLA and MBLAR possessed bla VIM-2 as a single cassette. DISCUSSION VIM-4-producing Klebsiella pneumoniae strains have been previously described from Tunisia [14], but to our knowledge this is the first report of VIM-2-producing P. aeruginosa from the African continent. Several surveys have been undertaken regarding the prevalence of MBL-producing isolates among CR P. aeruginosa, and varying percentages have been reported worldwide, ranging from 10% in Turkey [15] to up to 36% in Southeast Asia. [16] Our results are interesting in that this is the first study to show that 100% of the CR P. aeruginosa isolates from the Aga Khan University Hospital were MBL producers. It seems that other mechanisms of resistance to the carbapenems, such as
758 Clinical Microbiology and Infection, Volume 14 Number 8, August 2008 Dice (Tol 1.5%-1.5%) (H>0.0% S > 0.0%) [0.0%-100.0%] 55 60 65 70 75 80 85 90 95 100 CHR-01 CHR-02 CHR-03 CHR-04 AFC-03 AFC-04 AFC-07 AFC-02 MBLB AFC-12 AFC-13 AFC-15 AFC-01 AFC-14 MBLAR AFC-08 AFC-09 AFC-10 AFC-11 MBLA AFC-05 AFC-06 Fig. 1. Pulsed-field gel electrophoresis (PFGE) patterns and dendrogram of metallo-b-lactamase (MBL)-producing Pseudomonas aeruginosa strains isolated from Aga Khan University Hospital during 2006 2007. The dendrogram was constructed by cluster analysis using bionumerics software (version 3.5). Percentages of similarity are shown above the dendrogram. The MBLA, MBLAR and MBLB isolates formed separate clusters, each with >80% similarity among the PFGE profiles. The MBLAR isolates exhibited profiles >70% similar to those of the MBLA isolates, which indicates that both are related, although the significance is unknown. CHR-01, CHR-02, CHR-03, CHR-04: VIM-2-producing P. aeruginosa from the Calgary Health Region. AFC01 AFC15: VIM-2-producing P. aeruginosa from the Aga Khan University Hospital. AFC01, AFC02, AFC03, AFC04, AFC05, AFC07, AFC08, AFC12, AFC13, AFC14 and AFC15 were isolated from the intensive-care unit; AFC06, AFC09 AFC11 were isolated from general medical wards; and AFC10 was isolated from an outpatient. impermeability, which arises via the loss of the OprD porin [17], or the upregulation of an active efflux pump system present in the cytoplasmic membrane of these organisms [18], were absent at this tertiary-care centre in Nairobi. Molecular typing of VIM-2-producing P. aeruginosa showed that clonally related strains (the pan-resistant cluster MBLB with gene cassettes bla VIM-2 preceded by aaca4) were mostly present in the ICU and were responsible for nosocomial outbreaks in this unit during 2006 and 2007 (R. Revathi, unpublished data). Pan-resistant clusters of VIM-2-producing P. aeruginosa have also recently been described in Spain [19] and Korea [20]. Nearly half of the P. aeruginosa isolates from the ICU collected during the study period were MBL producers (33 67 (49%)); 21 of these (64%) were pan-resistant and belonged to cluster MBLB. Fifteen of the 33 (45%) MBL producers were isolated from respiratory secretions of septic patients, and 10 15 (67%) died as a result of the infection. According to the clinical data obtained from Kenya, these infected patients received antibiotic regimens combining piperacillin and amikacin, ceftazidime and amikacin, ciprofloxacin and amikacin, or ceftazidime and ciprofloxacin (colistin was not available in Kenya). This study supports data from Brazil that showed high mortality rates resulting from infections due to MBL-producing P. aeruginosa [21], especially in patients with ventilator-associated pneumonia [22]. This suggests that empirical therapy for suspected P. aeruginosa infections such as ventilator-associated pneumonia in the ICU of Aga Kahn University Hospital should include colistin.
Pitout et al. P. aeruginosa producing MBLs from Kenya 759 In summary, VIM-2-producing P. aeruginosa was identified at a tertiary-care centre in Nairobi, Kenya. A pan-resistant cluster was responsible for several nosocomial outbreaks in the ICU of this hospital, and continuous molecular surveillance is required to monitor the spread of this cluster. This study is the first report of VIM-2-producing P. aeruginosa from the African continent. ACKNOWLEDGEMENTS We would like to acknowledge K. Moore, and P. Le from Calgary Laboratory Services for their technical expertise. TRANSPARENCY DECLARATION None of the authors have professional, personal or financial conflicts of interest that would influence the conduct or reporting of this study. This work was partially supported by the DRESP2 (6th PCRD, LSHM-CT-2005-018705) contract with the European Commission. REFERENCES 1. Poirel L, Pitout JD, Nordmann P. Carbapenemases: molecular diversity and clinical consequences. Future Microbiol 2007; 2: 501 512. 2. Walsh TR, Toleman MA, Poirel L, Nordmann P. Metallo b-lactamases: the quiet before the storm? Clin Microbiol Rev 2005; 18: 306 325. 3. Gupta V. Metallo-b-lactamases in Pseudomonas aeruginosa and Acinetobacter species. Expert Opin Investig Drugs 2008; 17: 131 143. 4. Nordmann P, Poirel L. Emerging carbapenemases in Gram-negative aerobes. Clin Microbiol Infect 2002; 8: 321 331. 5. Laupland KB, Parkins MD, Church DL et al. Populationbased epidemiological study of infections caused by carbapenem-resistant Pseudomonas aeruginosa in the Calgary Health Region: importance of metallo-b-lactamase (MBL)-producing strains. J Infect Dis 2005; 192: 1606 1612. 6. National Committee for Clinical Laboratory Standards. Performance standards for antimicrobial susceptibility testing; fourteenth informational supplement M100-S14. Wayne, PA: NCCLS, 2004. 7. Pitout JD, Gregson DB, Poirel L et al. Detection of Pseudomonas aeruginosa producing metallo-b-lactamases in a large centralized laboratory. J Clin Microbiol 2005; 43: 3129 3135. 8. Pitout JD, Chow BL, Gregson DB et al. Molecular epidemiology of metallo-b-lactamase-producing Pseudomonas aeruginosa in the Calgary Health Region: emergence of VIM-2-producing isolates. J Clin Microbiol 2007; 45: 294 298. 9. Corvec S, Poirel L, Espaze E et al. Long-term evolution of a nosocomial outbreak of Pseudomonas aeruginosa producing VIM-2 metallo-enzyme. J Hosp Infect 2008; 68: 73 82. 10. Tenover FC, Arbeit RD, Goering RV et al. Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J Clin Microbiol 1995; 33: 2233 2239. 11. Paterson DL. The epidemiological profile of infections with multidrug-resistant Pseudomonas aeruginosa and Acinetobacter species. Clin Infect Dis 2006; 43 (suppl 2): S43 S48. 12. Poirel L, Naas T, Nicolas D et al. Characterization of VIM-2, a carbapenem-hydrolyzing metallo-b-lactamase and its plasmid- and integron-borne gene from a Pseudomonas aeruginosa clinical isolate in France. Antimicrob Agents Chemother 2000; 44: 891 897. 13. Aubron C, Poirel L, Fortineau N et al. Nosocomial spread of Pseudomonas aeruginosa isolates expressing the metallob-lactamase VIM-2 in a hematology unit of a French hospital. Microb Drug Resist 2005; 11: 254 259. 14. Ktari S, Arlet G, Mnif B et al. Emergence of multidrugresistant Klebsiella pneumoniae isolates producing VIM-4 metallo-b-lactamase, CTX-M-15 extended-spectrum b-lactamase, and CMY-4 AmpC b-lactamase in a Tunisian university hospital. Antimicrob Agents Chemother 2006; 50: 4198 4201. 15. Ozgumus OB, Caylan R, Tosun I et al. Molecular epidemiology of clinical Pseudomonas aeruginosa isolates carrying IMP-1 metallo-b-lactamase gene in a university hospital in Turkey. Microb Drug Resist 2007; 13: 191 198. 16. Huang YT, Chang SC, Lauderdale TL, Yang AJ, Wang JT. Molecular epidemiology of carbapenem-resistant Pseudomonas aeruginosa carrying metallo-b-lactamase genes in Taiwan. Diagn Microbiol Infect Dis 2007; 59: 211 216. 17. Livermore DM. Interplay of impermeability and chromosomal b-lactamase activity in imipenem-resistant Pseudomonas aeruginosa. Antimicrob Agents Chemother 1992; 36: 2046 2048. 18. Conejo MC, Garcia I, Martinez-Martinez L, Picabea L, Pascual A. Zinc eluted from siliconized latex urinary catheters decreases OprD expression, causing carbapenem resistance in Pseudomonas aeruginosa. Antimicrob Agents Chemother 2003; 47: 2313 2315. 19. Pena C, Suarez C, Tubau F et al. Nosocomial spread of Pseudomonas aeruginosa producing the metallo-b-lactamase VIM-2 in a Spanish hospital: clinical and epidemiological implications. Clin Microbiol Infect 2007; 13: 1026 1029. 20. Lee YC, Ahn BJ, Jin JS et al. Molecular characterization of Pseudomonas aeruginosa isolates resistant to all antimicrobial agents, but susceptible to colistin, in Daegu, Korea. J Microbiol 2007; 45: 358 363. 21. Zavascki AP, Barth AL, Goncalves AL et al. The influence of metallo-b-lactamase production on mortality in nosocomial Pseudomonas aeruginosa infections. J Antimicrob Chemother 2006; 58: 387 392. 22. Zavascki AP, Barth AL, Fernandes JF et al. Reappraisal of Pseudomonas aeruginosa hospital-acquired pneumonia mortality in the era of metallo-b-lactamase-mediated multidrug resistance: a prospective observational study. Crit Care 2006; 10: R114.