SEARCH WITHIN CONTENT
Citation Information : Polish Journal of Microbiology. Volume 70, Issue 1, Pages 107-116, DOI: https://doi.org/10.33073/pjm-2021-010
License : (CC-BY-NC-ND 4.0)
Received Date : 27-October-2020 / Accepted: 08-February-2021 / Published Online: 19-March-2021
Klebsiella pneumoniae is a critical multidrug-resistant (MDR) bacterium in humans responsible for numerous hospital infections linked to high morbidity and mortality since treatment options are limited (Navon-Venezia et al. 2017). K. pneumoniae from the family Enterobacteriaceae, occurs in the human and animal gastrointestinal tract microbiome. It is a commonly found opportunistic pathogen associated with the hospital environment and, overall, accountable for approximately a third of all Gram-negative infections. It has a role in extraintestinal infections, such as urinary tract infections, pneumonia, surgical site infections, cystitis, and life-threatening infections, including endocarditis and septicemia. It is also a significant cause of severe community-onset infections, such as necrotizing pneumonia, endogenous endophthalmitis, and pyogenic liver abscesses (Podschun and Ullmann 1998).
With the ever-growing antibiotic resistance, K. pneumoniae is a pathogen recognized for its antibiotic resistance; hence, it is categorized as an ESKAPE organism, besides other essential MDR pathogens (Boucher et al. 2009). The accumulation of ARGs by K. pneumoniae, by de novo mutations, is continuous under antibiotic selective pressure, and through the acquisition of plasmids and transferable genetic elements, it stimulates extensively drug-resistant (XDR) strains harboring a ‘super resistome’. In the past twenty years, many high-risk (HiR) MDR and XDR K. pneumoniae sequence types have appeared that exhibit great capacity of causing multicontinental outbreaks and continued global dissemination (Navon-Venezia et al. 2017).
Currently, the spread of carbapenem-resistant K. pneumoniae (CRKP) has become a severe problem in the molecular epidemiology of hospital infections.
Frequently, carbapenems serve as the last resort in the effective treatment of serious infections caused by multidrug-resistant bacteria. Enzymes that hydrolyze carbapenems, called carbapenemases, are the major cause of carbapenem resistance (Matsumura et al. 2017). The molecular classes A, B, and D carbapenemases are rapidly disseminated worldwide, challenging the treatment of Gram-negative infections (Nordmann and Poirel 2014). Recent reports have demonstrated that various carbapenem-hydrolyzing enzymes are disseminated worldwide in CRKP isolates. The fast evolution of carbapenem resistance quickly evolved in Enterobacteriaceae in the past decade and became a developing global threat. The majority of studies on antibiotic-resistant K. pneumoniae focus on characterizing carbapenemase producers (KPC, NDM, VIM, and OXA-48), various clonal groups or complexes (e.g., CG15, CG17, CG258, or CC147), and epidemic plasmids (IncA/C, IncFII, IncL/M, and IncN) that have been suggested to participate in their global expansion (Nordmann and Poirel 2014). Carbapenemase co-producers have been reported in distinct geographic locations: European countries (France, Germany, Greece, Italy, and Poland), Israel, the United States, China, and West Asia (Turkey) (Baraniak et al. 2011; Nordmann and Poirel 2014; Baraniak et al. 2015; Guo et al. 2016; Lee et al. 2016; Zautner et al. 2017; Bukavaz et al. 2018).
The majority of KPC-producing microorganisms also express β-lactamases and possess genes conferring resistance to other antimicrobials, i.e. aminoglycosides, fluoroquinolones, or co-trimoxazole (Nordmann and Poirel 2014). The resistance rates vary significantly across countries; MDR K. pneumoniae is endemic in Mediterranean countries, and Eastern and South-Western Europe. It stems from ESβL production in more than 50–60% strains, and non-susceptibility to third-generation cephalosporins, fluoroquinolones, and aminoglycosides (Navon-Venezia et al. 2017).
In 2011, the National Reference Center for Susceptibility Testing (NRCTS) and the KPC-PL Study Group published the first report from Poland that presented the molecular characteristics of K. pneumoniae producing KPC carbapenemases (Baraniak et al. 2011). Between 2011 and 2015, the bacteria caused 1,067 infection outbreaks; among them, 123 were caused by K. pneumoniae, and a higher number of outbreaks were reported from Masovian and Silesian voivodships (Baraniak et al. 2011). Poland belongs to the countries of the highest rate of K. pneumoniae isolates’ resistance to all groups of drugs subjected to monitoring, and these rates are twice as high as elsewhere in the European Union (EU)/European Economic Area (EEA). In these countries, resistant K. pneumoniae isolates consist on average 20.5% of all MDR multidrug-resistant strains (Bukavaz et al. 2018).
Given the abovementioned data and the increased frequency of isolation of CRKP strains from the hospital environment, we conducted a microbiological and molecular characterization of carbapenem-resistant K. pneumoniae isolates with emphasis on the antibiotic resistance profile, identification of ESβL genes, detection of carbapenemase genes, and the isolates’ genetic relationship.
Hospital settings and sample collection. The study was performed in the Upper-Silesian Medical Centre of the Silesian Medical University in Katowice (GCM), one of the largest multi-profile medical centres, and one of the largest hospitals in Poland. The hospital consists of 24 departments and treats over 160 thousand patients per year. Between February and August 2018, 15 non-duplicate CPKP isolates were collected from fecal samples of 14 patients admitted to three hospital wards characterized below. The Department of Neurology with the Stroke Sub-department (NR) receives 1,750 admissions per year and has 14 rooms with 44 beds; the Department of Internal Medicine and Rheumatology (REU) receives 1,935 admissions per year and has 12 rooms with 37 beds, and the Department of Anaesthesia and Intensive Care (OAIT) receives 1,750 admissions per year and has five rooms with ten beds (Table I). A total of 505 Enterobacteriaceae isolates were obtained from the patients in these wards over one-year.
Bacterial identification, antimicrobial susceptibility testing and phenotypic screening. Bacterial identification and preliminary susceptibility testing were performed using the automated VITEK® 2 Compact System (bioMérieux, France). The MICs of the 23 antimicrobial agents were evaluated with E-test strips (AB BIODISK, bioMérieux, France). MIC value results were interpreted according to the EUCAST breakpoints (EUCAST 2019). All isolates were screened phenotypically for the presence of KPC-, OXA-48-, and metallo-β-lactamases (MBL). Double-disc synergy tests (DDST) were carried out to confirm ESβL production (CLSI 2018) and to detect MBL production, as published previously (Matsumura et al. 2017). The phenotypic detection of KPC-producing isolates was assessed by an available protocol (Doi et al. 2008). According to the protocol, the boronic acid combined-disk tests using meropenem (10 μg) as the antibiotic substrate and 3-aminophenylboronic acid (3-APBA) (300 g, Sigma-Aldrich) as the inhibitor of KPC production was used. The detection of OXA-48 was performed in line with the EUCAST guidelines (Shaker et al. 2018). MDR and XDR were defined according to a standardized international document (Magiorakos et al. 2012).
Extended-spectrum β-lactamase genes (ESβLs) identification. Polymerase chain reaction (PCR), as described by Caltagirone et al. (2017) was used to determine blaSHV and blaTEM. CTX-Mplex PCR was used to detect blaCTX-M-type, blaCTX-M-1, and blaCTX-M-9 (Xu et al. 2005).
Carbapenemase genes detection. The existence of carbapenemase encoding genes (blaKPC-2, blaNDM-1, blaOXA-48, blaIMP, and blaVIM-1) was confirmed using the PCR method (Bukavaz et al. 2018).
PFGE. The genetic relatedness of isolates was investigated using the Pulsed Field Gel Electrophoresis (PFGE) method, following genomic DNA extraction and digestion with XbaI endonuclease (Fermentas, Lithuania), as described previously (Han et al. 2013). Salmonella enterica subsp. enterica serovar Braenderup strain H9812 (ATCC® BAA664™) and K. pneumoniae ATCC® BAA-1705™ were used as reference markers. PFGE banding patterns were compared using BioNumerics v.6.5 (Applied Maths, Belgium) software. The relatedness was determined by the unweighted pair group method using the average linkages (UPGMA), and the similarity of bands was calculated using the Dice coefficient.
MLST. Multilocus Sequence Typing (MLST) was conducted with the use of seven conserved housekeeping genes (gapA, infB, mdh, pgi, phoE, rpoB, and tonB) (Guo et al. 2016). A complete protocol of the MLST procedure, including the allelic profile and sequence type (ST) assignment techniques, is available in MLST databases.
Patients’ characteristics. Patients’ characteristics is presented in Table I. During seven months, the 15 K. pneumoniae isolates were collected from fourteen patients aged 65.5 years on average. The average duration of hospitalization was 37 days. All patients were found to be colonized by similar CRKP isolates. One patient developed a bloodstream infection (only blood isolate available) and was successfully treated with linezolid and colistin. Two of the improved patients were treated with carbapenem. Five patients died; their deaths were not associated with any particular etiology, though. CRKP strains were isolated at various stages of patients’ hospital stay. An average time from patient admission to the microorganism isolation was 16 days; in patient no. 3 the corresponding period was 64 days, while in patients nos. 1 and 5 – the period was only five days. None of the patients had traveled abroad shortly before their hospital admissions, so any travel experience could explain the colonization.
Epidemiologic investigation. The studied K. pneumoniae isolates’ epidemic curve exhibited a bimodal distribution of cases with two peaks separated by 148 days (Fig. 1). It could indicate double discrete periods (February-March 2018 and July-August 2018) with unobserved cases in January, April, May, and June. The NR ward was involved in these distinct periods. Following the index case, four new cases were registered in the OAIT, six cases in the NR, and three in the REU (Table I).
Antimicrobial resistance. The antimicrobial resistance profiles for the isolates are listed in Table II. We have classified nine (60%) strains as MDR and six (40%) as XDR. All isolates showed resistance to penicillins, cefalosporins, carbapenems (doripenem, ertapenem, meropenem), quinolone (ciprofloxacin), aminoglycosides (amikacin, gentamycin, netilmycin, tobramycin), and other antibiotics (aztreonam, trimethoprim/sulfamethoxazole, fosfomycin). Eleven isolates (73.3%) were sensitive to amikacin, two isolates (13.3%) were in the intermediate range with the MIC values equal to 12 μg/ml, and two isolates (13.3%) were resistant. Moreover, seven strains (42.8%) were resistant to tigecycline, five (35.7%) were in the intermediate range with the MIC values above 2 μg/ml. Interestingly, two isolates (13.3%) showed susceptibility to imipenem, and the resistance of one isolate (6.7%) to imipenem was within the intermediate range.
Phenotypic screening of carbapenemases and associated β-lactamases. In all K. pneumoniae isolates (n = 15, 100%), the ESβL mechanism was not found in phenotypic tests. Also, MBL production was not phenotypically detected in these isolates. However, all strains were positive for K. pneumoniae carbapenemase (KPC) production by the modified Hodge test (n= 15, 100%). In addition, phenotyping detection of carbapenem-resistant class D using the carbapenemase detection set with temocillin disc (30 μg) revealed that all isolates were positive for OXA-48 (n = 15, 100%).
β-lactamase genes detection. PCR amplification and sequencing analysis confirmed the presence of blaKPC-2, blaOXA-48, blaVIM-1 in all 15 of the K. pneumoniae isolates. In addition, the only identified blaCTX-M variant
in all 15 strains was blaCTX-M-15. The results are presented in Fig. 2. No isolate was found to be positive for blaSHV, blaTEM, blaNDM-1, and blaIMP.
Molecular epidemiology. Out of fifteen isolates concerned, all belonged to a single PFGE cluster (clone A) (Fig. 2).
Moreover, an MLST analysis allowed us to classify all isolates as the sequence type ST147 (allelic profile: 3-4-6-1-7-4-38) belonging to the CC147 clonal complex. It confirmed clonal relationships between the isolates.
The presented work stems from the problem of clonal dissemination of KPC-2, VIM-1, and OXA-48-producing K. pneumoniae ST147 noticed in one hospital in Katowice, the Silesian Voivodeship, Poland.
K. pneumoniae belongs to Gram-negative bacteria of the family Enterobacteriaceae. The pathogens can easily cause hazardous epidemic outbreaks and spread globally in the form of clones with increased virulence and epidemicity. These features are associated with the widespread presence of Enterobacteriaceae in humans (gastrointestinal carriage) and their great importance as frequent infection etiologic agent in facilities that carry out inpatient and outpatient treatment (Navon-Venezia et al. 2017). Studies that focus on K. pneumoniae are therefore desired, as they provide vital scientific grounds for the relevant action to be taken in the field of public health in respect of molecular epidemiology.
This study’s subject was a particular group of microorganisms classified as bacterial alarm agents (BCA) of particular virulence or resistance, i.e., carbapenemase-producing K. pneumoniae strains belonging to the order Enterobacterales. All K. pneumoniae isolates were derived from intestinal colonization cases in the patients studied. This work elaborated epidemiological opinions on the BCA epidemic threat occurrence in the multi-profile medical center in Katowice. It was also based on suspicion of an epidemic outbreak since we noticed clonal dissemination of several CRKP isolates and one case of symptomatic hospital infection with a blood CRKP strain. This strain was isolated in a given department 48 hours after the patient’s admission. The colonization of fourteen consecutive hospitalized patients with the same strain of the CRKP was also reported. CRKP isolates had unique properties such as the accumulation of numerous antibiotic resistance mechanisms and the capability to persist in a hospital environment. It resulted in special attention being paid to CRKP strains’ health care risk (Navon-Venezia et al. 2017). The genes encoding KPC, MBL, and OXA-48 are located on the so-called mobile genetic elements (plasmids, transposons) that allow effective spreading within bacterial populations. The genes encoding KPC carbapenemases (blaKPC) are located on the Tn4401 transposon located on plasmids with different types of replicons (IncF, IncL/M, ColE1, IncR, and IncX3). These plasmids show the ability to conjugate and propagate the blaKPC genes to new bacterial populations (Baraniak et al. 2011).
On the other hand, the most important families of the MBL acquired are the enzymes IMP and VIM occurring both in non-fermenting and intestinal bacilli. The blaIMP and blaVIM genes always exist as cassettes inserted into integrons. In turn, integrons can be located on transposons and move with them between DNA molecules (Izdebski et al. 2018b). ESβL enzymes exist mainly as acquired, plasmid-encoded β-lactamases. ESβL-encoding genes are often located on conjugation plasmids (IncFII, IncI), including those with a broad host range (IncA/C, IncL/M). It allows them to spread rapidly, also between strains belonging to different species. Frequently noted are encoded by blaCTX-M genes ESβL enzymes, active against cefotaxime CTX-M and localized on plasmids in Enterobacterales (Izdebski et al. 2018a; 2018b).
The isolates’ antibiotic susceptibility testing confirmed the β-lactamase antibiotics resistance profile typical for carbapenemase producers, as presented elsewhere (Zacharczuk et al. 2011; Guo et al. 2016; Izdebski et al. 2018a). All strains expressed high-level resistance to doripenem, ertapenem, meropenem, although they showed varying levels of resistance to imipenem. Ertapenem is the carbapenem that has been suggested as most suitable for detecting the presence of blaKPC, which was confirmed by high MIC values for ertapenem (> 32 μg/ml) in the present study. The presence of KPC does not always result in expression of in vitro resistance to imipenem or meropenem, as this carbapenemase shows limited efficiency of β-lactam ring hydrolysis in carbapenem particles. In this study, E-test showed resistance to ertapenem of 7/25804, 11/25808, 13/25810 strains, and susceptibility or medium susceptibility to imipenem (MICs 4–8 μg/ml), while PCR confirmed the presence of the blaKPC-2 gene in all three isolates. Imipenem and meropenem resistance of other strains might have resulted from decreased outer membrane permeability or other β-lactamases providing a synergistic effect with KPC (Protonotariou et al. 2018). The strains’ resistance to third-generation cephalosporins confirmed this hypothesis similarly to the various β-lactamases production by K. pneumoniae. The variability in CRKP strains’ carbapenem susceptibility illustrates the difficulties encountered while trying to identify strains producing carbapenemases only with the E-testing of resistance profiles. To reliably identify such strains, additional and more specific techniques should be applied (Xu et al. 2005; Magiorakos et al. 2012; Han et al. 2013; Caltagirone et al. 2017; Bukavaz et al. 2018).
The blaKPC-2, blaOXA-48, and blaVIM-1 carbapenemase encoding genes and an additional blaCTX-M-15 cefotaximase encoding genes have been found to coexist in all the strains concerned. Lee et al. (2016) found that K. pneumoniae can carry multiple β-lactamase genes in the same strain, which could be partly responsible for the pathogen’s selective success. All types of the bla genes were reported in combinations for this species (Lee et al. 2016). So far, K. pneumoniae that produce KPC-2 and KPC-3 carbapenemases have been frequently found in Europe, including Poland (Baraniak et al. 2011; Zacharczuk et al. 2011; Baraniak et al. 2015; Grundmann et al. 2017). On the other hand, Matsumura et al. (2017) presented the global dissemination of Enterobacteriaceae strains carrying the blaVIM genes. The VIM-producing Enterobacteriaceae are generally present in Europe, particularly in Greece, Spain, Hungary, and Italy (Matsumura et al. 2017). Significantly, an increasing problem with VIM-producing K. pneumoniae has been recorded in Greece, where the strains have accounted for the deaths of 48% of patients over the last two years (Protonotariou et al. 2018). In Poland, Izdebski et al. (2018b), collected one hundred and nineteen cases of VIM/IMP-positive Enterobacteriaceae in the period from 2006 to 2012, and showed many specific or entirely new features of these microorganisms, undoubtedly related to the properties of pathogens isolated in the abovementioned central-southern countries Europe, including several likely imported from abroad (e.g., from Greece) (Izdebski et al. 2018b).
The strains containing OXA-48 β-lactamases provide another example of rapid immigration of harmful microorganisms to Europe from their endemic regions, mostly eastern and southern Mediterranean countries (Egypt, Morocco, and Turkey). OXA-48-like-positive strains cause hospital outbreaks of epidemics in Belgium, France, Netherlands, Germany, Spain, and other countries (Nordmann and Poirel 2014; Grundmann et al. 2017). In the most recent report developed by OXA-48-PL Study Group monitoring the Enterobacteriaceae OXA-48-positive strains’ spread, the authors demonstrate that these strains have been isolated relatively rarely in Poland (Izdebski et al. 2018a). It was confirmed by other European and international research teams (Nordmann and Poirel 2014; Grundmann et al. 2017). Conjugative transfer of a specific plasmid group is regarded to be the central mechanism in the spread of OXA-48 among the Enterobacteriaceae populations (Nordmann and Poirel 2014).
The high MIC values (> 256 μg/ml) for cefotaxime may suggest that the strains produced CTX-M β-lactamases. Multiplex PCR and sequencing have confirmed blaCTX-M-15 presence in all isolates. The presence of various CTX-M β-lactamases in K. pneumoniae in Poland has been confirmed by other research teams (Baraniak et al. 2011; Izdebski et al. 2018a; 2018b). No blaSHV or blaTEM have been identified among the isolates. These results are in line with the persisting trend of diminished prevalence of hospital-associated K. pneumoniae strains producing SHV or TEM β-lactamases, which has also been noted by other scientists (Rodrigues et al. 2014). The type of ESβLs present in K. pneumoniae shifted in the 2000s, which caused outbreaks in hospitals due to the acquisition of plasmids and transposons encoding blaCTX-M-type ESβLs, consequently resulting in the predominance of CTX-M-producing strains (Calbo and Garau 2015).
Among non-β-lactam antibiotics, colistin has shown the highest activity towards the strains concerned. Similar results have been published by other authors (Zacharczuk et al. 2011). Among aminoglycosides, amikacin showed the highest activity. The vast majority of strains (92.8%) were susceptible to amikacin, unlike in China, where total K. pneumoniae resistance to amikacin was demonstrated (Guo et al. 2016). The other aminoglycosides (gentamicin, netilmicin, and tobramycin) were nonreactive to the CRKP isolates concerned, in line with data reported by other authors (Zacharczuk et al. 2011; Guo et al. 2016). On the other hand, Bukavaz et al. (2018) presented other aminoglycoside susceptibility profiles in CRKP strains. These discrepancies are most likely due to different quantities or origins of the strains concerned, though they might as well be related to particular hospitals’ or ward’s antibiotic policies.
The present analysis has shown a relatively high percentage of isolates resistant to tigecycline (42.8%), a synthetic minocycline analogue with a broad spectrum of activity. The antibiotic is successfully applied to treat infections caused by multi-resistant bacterial strains, including K. pneumoniae. During tigecycline treatment, Enterobacteriaceae representatives, including CRKP strains, may become resistant to the antibiotic, as evidenced by Pfarrell et al. (2018). They demonstrated cross-resistance between tigecycline and minocycline in resistant Enterobacteriaceae strains due to the MDR pumps, which remove drugs from cells. The distribution of acquired tigecycline resistance can vary depending on the geographic location or time; therefore, it is recommended to concern data on local resistance, particularly while treating severe infections (Pfarrell et al. 2018).
Genetic similarity of CRKP isolates was determined by comparing PFGE profiles (Han et al. 2013). This method has been widely used by European and non-European research teams that study epidemic outbreaks of K. pneumoniae and evidence global spread of this species’ epidemic clones (Guo et al. 2016; Bukavaz et al. 2018; Protonotariou et al. 2018). All K. pneumoniae strains analyzed in this study belonged to ST147 (CC147). Navon-Venezia et al. (2017) reported that ST147 K. pneumoniae clone spread globally. The involvement of this pandemic clone in the spread of CTX-M-15 and other carbapenemases in European, Asian, and Middle East countries was confirmed by others (Rodrigues et al. 2014; Guo et al. 2016; Zautner et al. 2017). The data found in the literature on K. pneumoniae clonal organization has been confirmed by research focused on the molecular epidemiology of infections due to antibiotic-resistant Enterobacteriaceae (Guo et al. 2016; Protonotariou et al. 2018).
Because of the clonal nature of the isolates studied, the clinical data about their origin was reviewed to examine whether the same clonal type was circulating in the three various hospital wards. The study focused mainly on stroke patients or patients suffering from cardiovascular complications who were transferred to multiple-bed hospital rooms in Stroke Recovery Units, which were the only places where they were in touch. Moreover, it is plausible that the neurology ward played a major role in disseminating the bacteria throughout the hospital. The spread of the pathogen isolates across three different hospital wards could imply that the personnel who can access all hospital wards freely, unlike the patients and their relatives, might have transmitted the infections. Scientific literature emphasizes the role of hospitalization (and its duration) in the process of bacterial colonization. The frequency of gastrointestinal colonization with K. pneumoniae strains in hospitalized patients was high in our study. Gastrointestinal colonization with CRKP is an important risk factor for infections with strains of the same phenotype due to the easy transfer of the genes and strains’ persistence in hospital environments, which hinders their eradication. This research evidences K. pneumoniae adaptation typical of endemic hospital strains, which are corroborated by the accumulation of locally spreading resistance mechanisms to the antibiotics most commonly used for therapeutic purposes.
Monika Brzychczy-Włoch https://orcid.org/0000-0002-7415-0154
The study was approved by the Bioethics Committee of the Jagiellonian University in Krakow, Poland (KBET/1072.6120.264.2019).
DO collected the data, performed the molecular analysis and drafted the manuscript; HK-C collected the isolates with clinical data, performed the hospital laboratory analysis and coordinated the microbiological analysis; MB consulted the cases and edited the manuscript; MBW supervised the research and analysis, coordinated and edited the manuscript.
The present work was financially supported by Jagiellonian University Medical College in Krakow (Poland), Grant No: K/ZDS/007829.
The authors are thankful to Karolina Winiewicz and Renata Wrona-Moczurad, the employees of the Department of Bacteriology of the GCM for their cooperation and assistance in conducting the study.
Gantt diagram depicting a timeline of patients’ admissions in relationship with the three wards and the length of stay in each ward. Epidemic curve based on the number of Abbreviations: OAIT-Department of Anaesthesia and Intensive Care, NR-Department of Neurology with the Stroke Subdepartment, REU-Department of Internal Medicine and Rheumatology.
Gantt diagram depicting a timeline of patients’ admissions in relationship with the three wards and the length of stay in each ward. Epidemic curve based on the number of
Abbreviations: OAIT-Department of Anaesthesia and Intensive Care, NR-Department of Neurology with the Stroke Subdepartment, REU-Department of Internal Medicine and Rheumatology.