Dissemination and Molecular Epidemiology of KPC-Producing Klebsiellapneumoniae Collected in Puerto Rico Medical Center Hospitals during a 1-Year Period

During a 2003-2004 PCR-based surveillance study conducted in 6 Puerto Rico Medical Center hospitals, 27/92 multi-beta-lactam-resistant Klebsiella pneumoniae strains were identiﬁed as carbapenemase (KPC) positive in 4 hospitals. The objectives of this study were to identify the KPC variants, their genetic relatedness, and any other beta-lactamases present. Susceptibility testing, pulsed ﬁeld gel electrophoresis (PFGE), isoelectric focusing, PCR, and DNA sequencing were performed. KPC variants -2, -3, -4, and -6 were identiﬁed. Additional beta-lactamases detected were TEM, DHA, OXA-9 and -30. Antimicrobial susceptibility to carbapenems varied depending on the KPC variant. Five PFGE genetically related groups were identiﬁed in 15 isolates and 12 unrelated types. PFGE proﬁles suggested that both clonal and horizontal transfer are contributing to the dissemination of these isolates among the various hospitals. Comparison of the 2003 and a 2009 surveillance studies showed a signiﬁcant increase in the KPC-positive K. pneumoniae isolates in the latter.

A retrospective antimicrobial resistance study performed at the Puerto Rico Medical Center suggested the presence of broad-spectrum beta-lactamases as one of the causes of the resistance problem [19]. During a 2003-2004 multibeta-lactam resistance surveillance study performed in this medical center, 92 of 285 (32%) unique and consecutive clinical isolates of K. pneumoniae were identified as multi-betalactam resistant. In this study, multi-beta-lactam resistance was defined as resistance to any of the carbapenems and/or two or more of the following antibiotics: ceftriaxone, cefotaxime, ceftazidime, cefepime, aztreonam, and piperacillintazobactam. PCR screening for the presence of familyspecific β-lactamase genes detected the KPC gene in 27 of the 92 multi-beta-lactam resistant (29%) or 10% of the 285 K. pneumoniae isolates. The objectives of this study were to identify the KPC variants, their genetic relatedness, and the presence of other β-lactamases in the 27 K. pneumoniae clinical isolates.

Susceptibility
Testing. Antimicrobial susceptibilities were determined by using commercial microdilution MIC panels (TREK Diagnostic System, Inc. Cleveland, OH, USA) following the manufacturer's instruction. The recent CLSI carbapenem susceptibility breakpoint changes for Enterobacteriaceae were utilized [18]. All isolates with an intermediate susceptibility to the antibiotics were considered resistant.

β-Lactamase PCR Screening.
The presence of familyspecific beta-lactamase genes were identified by PCR using panels of specific primers for the detection of the following genes: (1) plasmid-encoding AmpC β-lactamases (MOX, CMY, LAT, BIL, DHA, ACC, MIR, ACT, and FOX), (2) extended spectrum beta-lactamases (CTX-M groups I, II, III, and IV and TEM), (3) carbapenemase-hydrolyzing enzymes (KPC), and (4) metallo-β-lactamases (IMP and VIM). The oxacillinases OXA-30 and OXA-9 were used to try to identify some of the bands detected by isoelectric focusing (see below). The SHV-1 β-lactamase is chromosomally encoded in the majority of isolates of K. pneumonia; therefore, SHV primer was not used on the K. pneumoniae isolates. The bacterial DNA template, primers used, and the PCR conditions were performed as described previously [9,[20][21][22]. (IEF). IEF of sonicated crude cell extracts of the 27 KPC-positive isolates was performed as previously described [20,23] to determine the number of betalactamases and their isoelectric points (pIs), the capabilities of cloxacillin and clavulanate to inhibit the beta-lactamases, and the capability of the β-lactamase(s) to hydrolyze cefotaxime or imipenem.

Isoelectric Focusing
2.6. DNA Sequencing Analysis. PCR amplification of KPC full-length gene products was sequenced with primers KPCF1 and KPCR1, which flank the gene [9]. All PCR amplicons generated were sequenced independently and bidirectionally at least twice. Sequence alignment and analysis were performed online using the BLAST program (http:// www.ncbi.nlm.nih.gov/). PCR products were sequenced at the Creighton University Molecular Biology Core Facility using an ABI Prism 3100 Avant genetic analyzer (Applied Biosystems, Foster City, CA, USA).

Pulsed Field Gel Electrophoresis (PFGE).
To determine their genetic relatedness, the 27 K. pneumoniae isolates were analyzed by PFGE. For each isolate, DNA was prepared by in situ lysis of cells encased in agarose plugs, and digested with XbaI as previously described [24,25]. PFGE was performed using a Bio-Rad CHEF DR III System at 6 V/cm, 14 • C, 120 • C included angle, with switching from 5 to 15 s for 10 h, followed by switching from 15 to 60 s for 13 h. Images of ethidium-bromide stained gels were archived using a Bio-Rad Gel Doc 1000 System. PFGE profiles were compared using BioNumerics v 5.1. Isolates with 85% or greater similarity in the bands patterns were assigned to the same PFGE groups.

Comparison between 2003 and 2009 Surveillance Periods.
A comparison between the total number of multi-betalactam-resistant and KPC-positive K. pneumoniae collected from the Puerto Rico Medical Center for similar 6-month surveillance period for years 2003 (present study) and a recently published 2009 surveillance study (7) was performed.

Statistical
Analyses. The Two-tailed Fisher exact test or the Mixed Model Logistic Regression analyses were utilized to detect significant differences. A P value ≤0.05 was considered statistically significant.

Results
The 27 K. pneumoniae strains were isolated from four of the six Puerto Rico Medical Center hospitals (MC1, MC2, MC3, and MC4 and an unidentified hospital) with a total capacity of 629 beds. The types of clinical services offered by these hospitals include medicine, pediatrics, surgery, and trauma.  Table 1 shows the baseline epidemiological information of the patients with KPC-positive K. pneumoniae isolates. Twenty-one of 27 isolates were obtained from either hospital MC1 or MC3. A similar distribution of KPC-positive isolates was noted between females and males patients or the anatomical site of infection. Table 2 shows the KPC variants, isoelectric point (pI) value, and other beta-lactamases identified in the 27 isolates. Four KPC variants were identified: KPC-2 in 8 isolates, KPC-3 in 14 isolates, KPC-4 in 4 isolates, and the novel KPC-6 (GenBank accession number EU555534.1) in one isolate. The DNA sequence of this new variant showed a single point mutation at position 239 where a valine was substituted for a glycine. The IEF results indicated two pI values consistent with a KPC-like enzyme, a band with a pI of 7.65 identified as KPC-4 and a pI value of 6.7 for the rest of the variants including KPC-6. The pIs of the other betalactamases detected in the isolates ranged from 5.25 to 8.2, with the number of beta-lactamase bands per isolate varying from 2 to 7. Other beta-lactamases genes identified by PCR were the plasmid-encoded DHA-type AmpC in 4 isolates, OXA-30 in 5 isolates, OXA-9 in 6 isolates, and TEM in 26/27 isolates. None of the TEM beta-lactamases hydrolyzed cefotaxime on IEF and were not further characterized. Table 3 shows the in vitro susceptibility test results to selected antibiotics. The susceptibility to beta-lactam antibiotics demonstrated that 59% of the isolates were susceptible to cefepime followed by cefoxitin (37%), meropenem (37%), and imipenem (33%). None of the isolates were susceptible to ertapenem, ceftazidime, or aztreonam. For the non-betalactam antibiotics, the organisms were consistently susceptible to tigecycline (100%) and amikacin (89%). Susceptibility to polymyxins was not determined. Table 4 shows the susceptibility to carbapenems according to the KPC variant present in the K. pneumoniae isolates. All isolates were resistant to ertapenem irrespective of the KPC variant. Isolates with KPC-2 and -6 were resistant to all the carbapenem tested. The four isolates with KPC-4 were susceptible to imipenem and meropenem, while those with KPC-3 demonstrated variable susceptibility. Figure 1 shows the dendrogram and PFGE results to determine the genetic relatedness of the 27 KPC-positive K. pneumoniae isolates. A total of 17 distinct PFGE groups were    identified. Fifteen of the isolates were distributed between five distinct groups, and the remaining 12 isolates were unrelated to each other. Table 5 describes the characterization and distribution of the 15 K. pneumoniae isolates in the 5 PFGE-related groups. All isolates belonging to a related group possessed the same beta-lactamases and KPC variant. Hospital MC1 had 4/5 of the related groups followed by Hospital MC3 with 3/5. The other isolates from the remaining 3 groups were spread among the 4 hospitals, most of them distributed between Hospitals MC1 and MC3, while Hospital MC2 and MC4 with 1 related group each. Two hospitals each had a single unique related group, Hospital MC1 with group 11 and Hospital MC3 with group 14. Six of 11 (54%) of the related isolates were detected in the intensive care units (ICUs). Five of them were identified in Hospital MC1, representing 3/5 related groups. The distributions were as follows: 2 isolates from group 1, two from group 9, and two from group 11. Table 6 describes the distribution and characterization of the 12 PFGE-unrelated K. pneumoniae isolates. KPC variants and other beta-lactamases were also identified in these isolates. Hospitals MC1 and MC3 had most of the 12 unrelated isolates, with 4 isolates each, Hospital MC2 had 2, Hospital MC4 and the unidentified hospital 1 each. Eighty-three percent of these unrelated K. pneumoniae were identified in general hospitals wards and only 2 or 17% in the intensive care units. The difference in the distribution in either the ICU or general ward of the 5 PFGE-related groups and the 12 unrelated isolates was statistically significant (P ≤ 0.05). Table 7 shows the distribution of isolates in the four hospitals during the 1-year surveillance period according to related and unrelated PFGE groups and type of KPC variants. Hospitals MC1 and MC3 had the highest number of KPCpositive isolates with 13/27 and 8/27, respectively. KPC-2 and -3 were the most common variants observed and detected in all the hospitals. The KPC-2 variant was detected in PFGErelated group 15 and in five unrelated isolates (groups 2, 6, 7, 8, and 16). KPC-3 was identified in three related groups (1, 11, and 14) and in five unrelated isolates (groups 3, 5, 12, 13, and 17). KPC-4 was detected in related group 9 for two consecutive months in hospitals MC1, MC2, and MC3, and four months later, that same KPC variant was detected in a PFGE-unrelated isolate (group 10) in Hospital MC1. KPC-6 was identified once in hospital MC1 six months after starting the study. The movement of genetically related isolates into different hospitals was noted with three PFGE groups. PFGE Epidemiology Research International 5   9a (4) 15b (2) 3 (3) 15a (2) , 4 (6) 11a (3) , 2 (2) , 10 (4) 11a (3) 1a (3) , 1a (3) 1a (3) , 11a (3) 13 MC2 9a (4) 6 (2) 5 (3) 3 MC3 9a (4) 8 (2) 15a (2) 12 (3) 16 (2) 17 (3) , 14a (3) , 14b (3) 8 MC4 7 (2) 1b (3) 2 NR c 13 (3) 1 a Numbers followed by a letter represent the related PFGE groups (1, 9, 11, 14, and 15) and numbers alone represent the unrelated PFGE groups (2, 3, 4, 5, 6, 7, 8, 10, 12, 13, 16, and 17). b Type of KPC variant found in the isolates is shown in the superscript: KPC-2 (2) , KPC-3 (3) , KPC-4 (4) , and KPC-6 (6) . c NR: unidentified hospital (not reported).

Discussion
This paper represents the first molecular surveillance study of KPC and other beta-lactamases for K. pneumoniae strains at the Puerto Rico Medical Center hospitals. It is important to recognize that patients admitted to the PRMC represent, for the most part, referrals from all over the island of complicated or seriously ill patients. Most of these patients have been previously hospitalized at other institutions and subjected to different therapies including parenteral antibiotics. Because of the complicated nature of their conditions, many patients admitted have prolonged hospitalizations, several invasive procedures, and extensive antibiotic treatments. All these factors may contribute to the patients' acquisition of multidrug-resistant bacteria.
Four KPC variants were detected but, as expected, KPC-2 and KPC-3 were the most prevalent [1]. KPC-4 with a pI value of 7.65 was identified in 4 isolates belonging to PFGE groups 9 and 10. This variant, initially detected in Enterobacter sp. from Scotland [26] is, to our knowledge, identified for the first time in K. pneumoniae. A new variant, KPC-6, was identified in one isolate. The detection of multiple KPC variants in our isolates may represent random mutational events or introduction of the genes from external sources. Additional beta-lactamases identified in these isolates are OXA-30, OXA-9, TEM, and plasmid-encoded AmpC-DHA. The presence of multiple beta-lactamases in KPC-positive isolates has been previously reported [1,3,14,27].
The in vitro antimicrobial resistance to the beta-lactam antibiotics, fluoroquinolones, trimethoprim sulfa, and gentamicin was observed suggesting the presence of multiple mechanisms of antibiotics resistance. Only amikacin and tigecycline demonstrated good to excellent in vitro antimicrobial activity against these isolates; unfortunately, susceptibility to the polymyxins were not performed. These data are in agreement with previous reports showing the limited therapeutic options available for the treatment of infections caused by KPC-positive bacteria [1,28].
All K. pneumoniae isolates were resistant to ertapenem, but nearly one-third were susceptible to imipenem and/or meropenem, even though the recent CLSI carbapenem breakpoints were used [18]. Carbapenem susceptible isolates harbored either the KPC-3 or -4 variant. These results may be explained on the basis of several interacting bacterial factors, such as (1) intrinsic hydrolytic activities of the different KPC variants, (2) the genetic background and control of carbapenemase production, and (3) the presence or absence of other broad-spectrum beta-lactamases or porin mutations [11,14,26,27,[29][30][31]. Although the number of isolates is small, our data is in agreement with previous reports which suggest that a K. pneumoniae resistant to ertapenem should alert the laboratory staff and the clinician of the possible presence of a KPC-positive isolate and the need for further testing [2,17,32].
The PFGE data demonstrated a combination of bla KPC movements among different strains, as well as the spread of specific related groups to different sites. This implies that one or more of the following mechanisms may be responsible for the rapid intra-and interhospital dissemination of these isolates: (1) colonized human contacts (health care professionals, patients, and other individuals), (2) fomites or medical equipment as a vehicle for transmission, (3) the extensive use of broad-spectrum antibiotics creating selective pressure, and (4) the presence of the bla KPC within mobile genetic elements. The comparative six month data for the years 2003 and 2009 demonstrates a significant increase in the relative and absolute numbers of KPC-positive K. pneumoniae for all 4 hospitals. This may represent a combination of multiple factors such as, among others, a failure of either the hospital personnel and/or medical staff to follow the established infection control policies, increase in personnel and medical staff transit between hospitals, inability to identify an isolate promptly admitted patients colonized or infected with multidrug-resistant organisms acquired in other institutions, lack of an effective alert system between the laboratory, hospitals infection control programs, and/or the inappropriate long standing use of broad-spectrum antibiotics. The present study does not identify the relative importance of these factors or the impact of understaffing common to all hospitals. It is imperative that hospitals strictly enforce the infection control guidelines, appropriate antibiotic utilization practices, and the Center for Disease Control and Prevention recommendations for the detection and control of carbapenem-resistant Gram-negative bacilli to reduce the spread of these isolates as they have been associated with significant mortality [33][34][35].