Summary

Introduction: In this study, we aimed to determine the epidemiological properties of metallo-beta-lactamase-producing Pseudomonas aeruginosa (MBL-PA) isolates and to investigate the relationship between the presence of MBL-PA and patient morbidity and mortality.

Materials and Methods: The study included carbapenem-resistant P. aeruginosa isolates recovered from various clinical specimens of 334 patients in Karadeniz Technical University Faculty of Medicine Hospital, a 900-bed university hospital in Trabzon, Turkey. MBL-related carbapenem-resistant PA strains were phenotypically investigated using the Modified Hodge test and the imipenem/imipenem-ethylene diamine tetra acetic acid combined disc tests. Multiplex polymerase chain reaction was used to investigate the presence of bla_IMP, bla_VIM, bla_GIM, bla_SIM, and bla_SPM genes, which are responsible for MBL production. Clonal relationships among MBL-PA isolates were analyzed by pulsed-field gel electrophoresis. The patients" hospital records were retrospectively examined. Various demographic and clinical characteristics were evaluated in relation to MBL-PA.

Results: Thirty-two (9.6%) of the carbapenem-resistant PA isolates were found to carry bla_VIM and/or bla_IMP, with three strains harboring both blaVIM and blaIMP. MBL-PA isolates were more resistant to aminoglycosides and quinolones. Eight Verona integron-encoded metallo-beta-lactamase-type MBL-PA isolates were found to be identical in adults, while several clonally-related clusters were observed among MBL-PA isolates in both the pediatric and adult inpatients. Compared to non-MBL carbapenem-resistant PA, the risk factors evaluated were found to have no association with MBL-PA. In addition, there was no statistically significant difference in mortality between patients from whom MBL-PA or non-MBL-PA was isolated.

Conclusion: Although MBL-PA has been implicated in various healthcare-related outbreaks, no specific risk factor has been identified in association with MBL-PA isolation. To our knowledge, this is the first study in Turkey to detect P. aeruginosa isolates carrying both blaVIM and blaIMP.

Introduction

There is a need to assess the current epidemiological status of MBL-PA in the local setting and to delineate the mechanisms which lead to resistance. By precisely determining the clinical and molecular characteristics of the isolates involved, we may better understand their epidemiology, since these characteristics form the basis for effective epidemiological surveillance. The surveillance in turn helps us to gauge the effectiveness of infection control measures and to guide correctly targeted and timely treatment measures[1-3]. According to the hitherto limited research conducted in this field in Turkey, VIM-type MBLs are the most commonly reported MBL types in P. aeruginosa isolates[3-8]. More comprehensive research on this subject is required to reveal the absolute epidemiology of resistance and the clinical and molecular characteristics of MBL-PA in Turkey[3].

In order to assess the molecular epidemiology in terms of dissemination of MBL-PA isolates and the clinical characteristics related to MBLs in our hospital, we aimed to detect MBL-PA isolates, demonstrate the molecular epidemiology of MBL-PA, and determine any clinical risk factors associated with MBL-PA and the clinical outcomes of MBL-PA isolation among inpatients.

Methods

Medical records of the patients were retrospectively reviewed. Hence, owing to the retrospective nature of this part of the study, it was not deemed necessary to obtain written consent from the patients. Due to inadequacy of the patients" hospital records, it could not be ascertained whether the P. aeruginosa isolates represented colonization or were the causative agents of an infection.

Phenotypic investigation of MBL producers was performed by both the MHT and the imipenem/imipenem-EDTA CDT. Multiplex polymerase chain reaction (PCR) was performed to investigate carriage of the bla_IMP, bla_VIM, bla_GIM, bla_SIM, and bla_SPM genes. The molecular epidemiology and microbiological characteristics of the MBL-PA isolates were evaluated in two groups: a pediatric inpatient group and an adult inpatient group.The genetic relatedness of carbapenem-resistant MBL-PA isolates was investigated by pulsed-field gel electrophoresis (PFGE) for each group[13-18].

Retrospective Case-Control Study
A retrospective case-control study was performed to evaluate risk factors related to MBL-PA. The risk factors analyzed were: age, sex, underlying diseases, comorbidities, source of the carbapenem-resistant P. aeruginosa isolate, previous surgical operations, invasive device usage (central venous catheter, endotracheal tube, and urinary catheter), mechanical ventilation, immunosuppression lasting longer than 14 days (e.g. chemotherapy, corticoids), and antimicrobial usage (for at least 48 h over the preceding 14 days). A total of 158 patients (from a potential 334 patients) with available hospital records were retrospectively reviewed from the pediatric (n=23) and adult (n=135) inpatient groups. Outpatients (n=42) were excluded from the risk assessment and the clinical outcome evaluation. Patients whose samples yielded MBL-PA were defined as the case group, while the patients with non-MBL-PA isolates were defined as the control group. Clinical outcome was assessed based on the length of hospital stay and whether the final outcome was discharge or death. MB-PA-related mortality was defined as death occurring within ten days of MBL-PA being isolated[9, 10].

Bacterial Identification and Susceptibility Tests
Identification and susceptibility testing of the isolates was performed using standard laboratory methods and Phoenix NMIC/ID-55 panels (Becton Dickinson Bioscience automatic system; USA) in accordance with the manufacturer"s instructions. P. aeruginosa ATCC 27853 was used as a quality control strain. Antimicrobial susceptibility test results were interpreted according to the recommendations of the Clinical and Laboratory Standards Institute[11].

Modified Hodge Test
A suspension of Escherichia coli ATCC 25922 was inoculated on a Mueller-Hinton agar after the density was adjusted to McFarland 0.5 standard, after which a 10-μg imipenem disc (Oxoid Thermo Fisher, UK) was placed at the center of the agar plate. The test isolates, positive control strain (Klebsiella pneumoniae ATCC BAA-1705), and negative control strain (K. pneumoniae ATCC BAA-1706) were streaked in a straight line from the edge of the disc to the edge of the plate, in different directions for each isolate. In case of E. coli ATCC 25922 growth on the test isolate streak line towards the imipenem disc after overnight incubation, the so-called "clover leaf" pattern, the test isolate was interpreted as MHT-positive[11].

Imipenem/Imipenem-Ethylene Diamine Tetra Acetic Acid Combined Disc Test
Suspensions of the test isolates and control strains were inoculated on Mueller-Hinton agar after adjusting the density to McFarland 0.5 standard. Two 10-μg imipenem discs (Oxoid Thermo Fisher, UK) were placed on the agar plate with a distance of 20 mm between their centers. Ten μL of 0.5 M EDTA was added onto one of the imipenem discs. After overnight incubation, the discs were examined. If the zone of inhibition surrounding the disc impregnated with both imipenem and EDTA was at least 7 mm greater in size than the zone around the disc containing imipenem alone, the isolate was considered MBL-positive[12]. IMP-positive 587585 P. aeruginosa and VIM-positive 670448 P. aeruginosa isolates, which Ozgumus et al.[5] reported to be producers of MBL, were used as positive controls, while P. aeruginosa ATCC 27853 was used as a negative control[5].

Polymerase Chain Reaction Detection of Metallo-beta-lactamase Genes
Bacterial DNA was extracted by the boiling method[13]. Polymerase chain reaction was performed using a bacterial DNA template together with the specific primers listed in Table 1[14, 15]. IMP-positive 587585 P. aeruginosa and VIM-positive 670448 P. aeruginosa isolates were used as positive controls. P. aeruginosa ATCC 27853 and distilled water were used as negative controls[5]. For the bla_IMP (A), bla_VIM (A), bla_GIM, bla_SIM, and bla_SPM genes, the PCR amplification conditions were as follows: initial DNA denaturation at 94 °C for 5 min, followed by 35 cycles of denaturation at 94 °C for 20 sec, annealing at 53 °C for 45 sec, and extension at 72 °C for 30 sec, followed by final extension at 72 °C for 6 min[16]. For the bla_IMP (B) and bla_VIM (B) genes, PCR amplification was done using the following conditions: initial DNA denaturation at 94 °C for 5 min, followed by 35 cycles of denaturation at 94 °C for 25 sec, annealing at 57 °C for 40 sec, and extension at 72 °C for 50 sec, followed by final extension at 72 °C for 6 min[15].

Table 1: List of primers used in this study[14, 15]

Pulsed-field Gel Electrophoresis
Chromosomal DNA was prepared as previously described. SpeI (Promega Corp., USA) was used to digest the genomic DNA. Lambda phage concatemers (Bio-Rad Laboratories, USA) were used as a size marker. Electrophoresis was carried out under the following conditions using the CHEF-DR III system (Bio-Rad Laboratories, USA): an initial switch time of 5 seconds, final switch time of 35 seconds, gradient of 6 V/cm, and included angle of 120° for a 24-hour run[17]. Bio-Rad Gel Doc System (Bio-Rad Laboratories, USA) was utilized to document the PGFE patterns, and the clonal relationship among isolates was analyzed by Molecular Analyst Software (Bio-Rad Laboratories, USA) using the Dice similarity coefficient. Isolates with ≥95% genetic similarity in PFGE profiles were defined as being from the same clone, while clusters were defined as DNA patterns with ≥85% similarity[18].

Statistical Analyses
The SPSS 13.0 program (SPSS Inc., USA) was used for all statistical analyses. The Kolmogorov-Smirnov (K-S) test was used to determine normality. For binomial comparisons of numerical data, Student's t-test was used for normal distributions and the Mann-Whitney U test was used for non-normal distributions. The independent-samples t-test was used to compare means of the data. The χ2 test was used for qualitative comparisons. The Kaplan-Meier test was used for survival analysis. Statistical tests with p<0.05 were considered statistically significant.

Results

Table 2: Evaluation of Modified Hodge test and imipenem/imipenem-ethylene diamine tetra acetic acid combined disc test for detection of metallo-beta-lactamase-producing Pseudomonas aeruginosa isolates

Table 3: Polymerase chain reaction detection of blaVIM and blaIMP genes regions using different specific primers

Microbiological Characteristics of Carbapenem-resistant P. aeruginosa Isolates
Carbapenem-resistant P. aeruginosa organisms were isolated from several clinical specimens: 165 (49.4%) were isolated from respiratory tract samples, 63 (18.9%) from genitourinary system samples, 59 (17.6%) from skin and soft tissue samples, 28 (8.4%) from blood samples, and 19 (5.7%) from other samples.

Carbapenemase production was detected by MHT in 21 (6.3%) of the 334 P. aeruginosa isolates. MBL production was found in 109 (32.6%) isolates with CDT (Table 2). Among 334 carbapenem-resistant P. aeruginosa isolates, 32 (9.6%) were positive for bla_VIM or/and bla_IMP, and 3 of these isolates were harboring both bla_VIM and bla_IMP (Table 3).

Eight of the MBL-PA were isolated from pediatric inpatients. Three of them were bla_VIM-positive, four were bla_IMP-positive, and one was both bla_VIM- and bla_IMP-positive. Among P. aeruginosa isolates found to be MBL-positive, 17 were isolated from the adult inpatient group, of which 13 were VIM-positive only and four were IMP-positive only. Seven of the MBL-PA were isolated from outpatients, of which four were bla_VIM-positive, one was bla_IMP-positive, and two were found to be positive for both bla_VIM and bla_IMP.

Antibiotic resistance patterns of the carbapenem-resistant P. aeruginosa isolates are summarized in Graphic 1. Compared to non-MBL-PA, MBL-PA isolates showed higher rates of resistance for antibacterial agents such as piperacillin, ceftazidime, cefepime, piperacillin-tazobactam, aztreonam, gentamicin, amikacin, and ciprofloxacin.

Graphic 1: Evaluation of antibiotic resistance rates (%) of metallo-beta-lactamase-producing Pseudomonas aeruginosa and non-metallo-beta-lactamase-producing Pseudomonas aeruginosa MBL-PA: Metallo-beta-lactamase-producing Pseudomonas aeruginosa

Molecular Epidemiology of Metallo-beta-lactamase-producing P. aeruginosa Isolates in the Pediatric Inpatient Group
All the bla_VIM or/and bla_IMP-harboring PA strains (n=8) isolated from the pediatric inpatient group were compared by PFGE. Three different pulsotypes were obtained among the isolates of PA harboring bla_VIM (n=3) and PA harboring bla_VIM and bla_IMP (n=1), while five different pulsotypes were obtained from isolates of PA harboring blaIMP (n=4) and PA harboring blaVIM and blaIMP (n=1) (Table 4). Three particular bla_VIM-harboring PA isolates (PA206, PA14, and PA19) were considered to form a cluster, with similarity over 85%. Furthermore, two of them (PA14, and PA19) were effectively identical, with similarity over 95%. With similarity over 85%, two bla_IMP-harboring PA isolates (PA31, and PA35) were considered as a cluster, whereas a bla_IMP-harboring PA isolate (PA140) and the bla_VIM and bla_IMP-harboring PA isolate (PA146) were defined as another cluster.

Table 4: Pulsed-field gel electrophoresis patterns of the bla_VIM-positive Pseudomonas aeruginosa and bla_IMP-positive Pseudomonas aeruginosa isolates in the pediatric inpatient group

Molecular Epidemiology of Metallo-beta-lactamase-producing P. aeruginosa Isolates in the Adult Inpatient Group
All bla_VIM or bla_IMP-harboring PA isolates (n=17) isolated from adult inpatients were compared by PFGE (Table 5). Six different pulsotypes were obtained in blaVIM-harboring PA (n=13), while four different pulsotypes were obtained in blaIMP-harboring PA (=4). Eleven bla_VIM-harboring PA isolates (PA230, PA237, PA250, PA282, PA238, PA254, PA43, PA328, PA241, and PA236) were considered a cluster, with over 85% similarity. In addition, eight of those strains (PA282, PA238, PA298, PA254, PA43, PA328, PA241, and PA236) were identical, with similarity exceeding 95%. Pulsotypes obtained from bla_IMP-harboring PA isolates (PA166, PA139, PA145, and PA136) were less than 75% similar.

Table 5: Pulsed-field gel electrophoresis patterns of the bla_VIM-positive Pseudomonas aeruginosa and bla_IMP-positive Pseudomonas aeruginosa isolates in the adult inpatient group

Clinical Characteristics of Metallo-beta-lactamase-producing P. aeruginosa Isolates
From a total of 334, 158 patients" hospital records were available. Of these, MBL-PA was isolated from 16 patients and non-MBL-PA was isolated from 142 (Figure 1). No statistical difference was found between patients with MBL-PA and those with non-MBL-PA in terms of age, gender, stay in the intensive care unit (ICU), length of stay in hospital before infection, site of infection, whether or not the infection was polymicrobial, presence of any other accompanying infection, underlying disease, or invasive procedure, history of surgery or trauma, hospitalization within the last 30 days, length of hospital stay, administration of antibiotherapy, or immunosuppressive therapy (Table 6). Nine (43.8%) of the patients with MBL-PA died, while 70 (50.7%) of the patients with non-MBL-PA died. No significant difference was found between patients with MBL-PA and non-MBL-PA in terms of mortality (Log Rank: 0.536, p=0.384).

Figure 1: Summary of study findings for MBL-PA and non-MBL-PA

aOf 221 adult inpatients, hospital records were available for 135 patients, 11 of whom had MBL-PA isolates and 124 with non-MBL-PA isolates. bOf 71 pediatric inpatients, hospital records were available for 23 patients, 5 of whom had MBL-PA isolates and 18 with non-MBL-PA isolates. cA total of 42 outpatients were excluded from the PFGE analysis, risk assessment, and clinical outcome evaluation.

MBL-PA: Metallo-beta-lactamase-producing Pseudomonas aeruginosa, PFGE: Pulsed-field gel electrophoresis, IMP: P. aeruginosa include imipenemase, VIM: Verona integron-encoded metallo-beta-lactamase

Table 6: Statistical analysis of risk factors for metallo-beta-lactamase-producing Pseudomonas aeruginosa infections

Table 6: Continued

Table 6: Continued

Discussion

Carbapenemases are responsible for an important part of carbapenem resistance and are a source of concern worldwide. Various types of carbapenemases have been increasingly reported in P. aeruginosa strains over the years[2, 3, 19, 20]. The distribution of carbapenemases varies according to geographical region, but the most prevalent MBL enzyme types are VIM and IMP[2]. Verona integron-encoded metallo-beta-lactamase-type MBLs in P. aeruginosa have been the most commonly reported carbapenemases in Turkey to date, whereas P. aeruginosa producing IMP-type MBLs have only been reported from Trabzon and Muş[3-8]. Include imipenemase-type MBLs have not yet been reported in Enterobacteriaceae or Acinetobacter baumannii isolates in Turkey[21-25]. In the present study, 32 MBL-PA were detected among 334 carbapenem resistant P. aeruginosa isolates (9.6%), and bla_VIM was more prevalent than bla_IMP. Additionally, considering that our study was conducted at the same hospital from which the first IMP-type MBL was detected, additional studies should be performed in different regions of the country in order to ascertain whether the VIM-type MBLs are limited to these regions[5-8]. Our study appears to be the first from Turkey reporting PA isolates with both bla_VIM and bla_IMP carriage.

Compared with non-MBL-PA isolates, MBL-PA isolates were found to be more resistant to aminoglycosides, quinolones, and even aztreonam in this study. Although aztreonam is not a substrate for MBLs, it is often ineffective against these strains due to additional mechanisms of resistance[1, 26].

Additional resistance mechanisms in MBL-PA commonly include cephalosporinase, efflux pumps, or low intrinsic outer membrane permeability. Furthermore, MBL genes and genes encoding other antibiotic resistance determinants may be located on the same plasmids[27]. Therefore, additional resistance mechanisms could be acquired simultaneously with MBL genes and lead to resistance to antibiotics other than beta-lactams such as aminoglycosides and quinolones[1, 26]. Depending on the factors listed, it is possible to detect high resistance rates and multidrug resistance in MBL-PA isolates, as demonstrated by this study.

PFGE analysis demonstrated several clonally-related genotype clusters among the VIM-type MBL-PA and IMP-type MBL-PA isolates in the pediatric inpatient group, while eight of the VIM-type MBL-PA isolates from the adult inpatient group were identical. These findings indicate that cross-transmission is an important mechanism for dissemination of MBL-PA, resulting in multidrug resistance in P. aeruginosa isolates.

Potential risk factors identified for infection or colonization with MBL-PA include ICU stay, long-term hospitalization, use of indwelling urinary catheters, urinary tract diseases, hemodialysis, and hospitalization within the preceding year, administration of antineoplastic agents or corticosteroids, fluoroquinolone usage, and long-term antibiotic use (especially beta-lactams)[9, 28-31]. Nonetheless, we found no statistically significant difference between MBL-PA and non-MBL-PA in terms of demographic features, stay in ICU, length of hospital stay before isolation, sample source, underlying disease, invasive procedure, or history of antibiotic therapy, hospitalization, surgery, trauma, or immunosuppressive therapy within the last 30 days. While some researchers have reported that MBL-PA infections result in a higher mortality rate than non-MBL-PA infection, others have found no association between MBL-PA and mortality[28, 30, 31]. We also found no statistically significant differences in terms of mortality. This finding may be because of our limited data about the patients. Differences of mortality rates may also be related to the virulence properties of the infecting strains.

One of the limitations of this study is that because of its retrospective design, records were not available for all patients. In addition, the MBL-PA groups were significantly smaller in number than the non-MBL-PA groups due to the prevalence of MBLs.

Conclusion

Ethics
Ethics Committee Approval: The study protocol was approved by Karadeniz Technical University Faculty of Medicine Ethic Council (decision date-number: 2011/7-2).

Informed Consent: Written informed consent of the patient was not obtained due to the retrospective nature of this study but hospital records were reviewed by the approval of the hospital management.

Peer-review: Externally peer-reviewed.

Authorship Contributions
Surgical and Medical Practices: Y.B., G.B., Concept: G.B., F.A., Design: Y.B., G.B., Data Collection or Processing: Y.B., N.K., Analysis or Interpretation: G.B., İ.T., Literature Search: Y.B., G.B., Writing: Y.B.

Conflict of Interest: No conflict of interest was declared by the authors.

Financial Disclosure: The authors declared that this study received no financial support.