Background

The genus Providencia comprises part of the natural human gut flora but may also cause infections, including travelers’ diarrhea, urinary tract infections, and other nosocomial infections [1]. Treatment of these infections is challenging because Providencia rettgeri strains are intrinsically resistant to many antimicrobials including ampicillin, first generation cephalosporins, polymyxins and tigecycline [2]. Furthermore, in recent years P. rettgeri has become increasingly important because of the emergence of carbapenemase-producing strains [3, 4]. Carbapenemases are enzymes known to hydrolase almost all types of β-lactams [5]. The New Delhi metallo-β-lactamase (NDM-1) has been firstly identified in 2009 in a Swedish patient who had been previously hospitalized in New Delhi, India [6]. The first occurrence of NDM-1 producers was reported in clinical isolates of P. rettgeri in Israel in 2013 [7]. Since then, other cases have been reported in Mexico, Brazil, Argentina, Ecuador, Canada, and Nepal [3, 4, 8,9,10,11,12,13].

PER-1 enzyme is belong to class A extended-spectrum β-lactamases (ESBLs) and firstly discovered in a plasmid of Pseudomonas aeruginosa in France [14]. Later, it has also found among several Gram-negative species including Acinetobacter baumannii, Salmonella enterica serovar Typhimurium, and also in P. rettgeri [15, 16]. PER-1 is widely spread in Turkey, however, high prevalence of PER-1 ESBL in A. baumannii has been reported in Korea [17].

Here, we report the first outbreak of multidrug-resistant P. rettgeri strain co-producing NDM-1 and PER-1 in Korea.

Materials and methods

Patients and bacterial isolates

From January 2016 to January 2017, a total of eight P. rettgeri isolates from eight patients were included in this study. Bacterial identification was done with a Vitek-MS (bioMérieux, Marcy I’Etoile, France). Medical records of the patients were retrospectively reviewed. This study protocol was approved by the hospital institutional review board.

Antimicrobial susceptibility testing

Minimum inhibitory concentrations (MICs) for cefotetan, cefotaxime, ceftazidime, cefepime, ertapenem, imipenem, meropenem, aztreonam, amikacin, ciprofloxacin, gentamicin, and tigecycline were determined using Etest strips (bioMérieux) on the Mueller–Hinton agar (Becton–Dickinson, Sparks, MD, USA). Colistin MIC was determined by broth microdilution. When available, antimicrobial susceptibility was interpreted based on the Clinical and Laboratory Standards Institute (CLSI) guideline [18]. For tigecycline and colistin, the European Committee for Antimicrobial Susceptibility Testing (EUCAST) criteria were used [19].

Detection of β-lactamase genes

The carbapenemase genes and ESBL genes were detected using specific PCR primers (Table 1) [20,21,22,23,24,25,26,27]. Amplified products were directly sequenced on the ABI 3730xl automatic sequencer (Applied Biosystems, Foster City, CA, USA) using the same primer pair. The sequences obtained were compared to those in GenBank (www.ncbi.nlm.nih.gov/GenBank) using the BLAST program (www.ncbi.nlm.nih.gov/BLAST/).

Table 1 Primers used in this study for identifying antimicrobial resistance genes

Pulsed-field gel electrophoresis

The bacterial genetic relatedness was evaluated by Pulsed-field gel electrophoresis (PFGE). Genomic DNA was digested with SfiI enzyme, and DNA fragments were separated on a CHEF-DRII System (Bio-Rad, Hercules, CA, USA). A lambda ladder (Bio-Rad) was used as a DNA size marker. The band patterns were analyzed using UVIband/Map software (UVItech Ltd., Cambridge, UK) and the dendrograms were generated based on the unweighted pair group method using arithmetic averages from the Dice coefficient. Isolates that exhibited a PFGE profile with more than 90% similarity (pulsotype) were considered as closely related strains.

Results

The characteristics of these patients and antimicrobial susceptibility patterns of P. rettgeri isolates were summarized in Table 2. In total, eight P. rettgeri isolates were recovered from urine samples of eight patients admitted in a surgical intensive care unit (SICU). All patients were admitted to a SICU from hospitalization and had a urinary catheter. The median days of the SICU stay before P. rettgeri isolation was 21.5 days (range, 8–38 days) (Fig. 1). All patients except one (P5) were recovered and discharged during the outbreak. A patient (P5) died following Enterococcus faecalis bacteremia. All P. rettgeri isolates showed similar antibiogram with high MIC levels to various classes of antimicrobial agents tested (cefotetan, cefotaxime, ceftazidime, cefepime, azteronam, meropenem, ertapenem, ciprofloxacin, amikacin, and tigecycline). Imipenem MICs were 0.5–4 μg/mL (6/8 susceptible isolates, 1/8 intermediate isolate, and 1/8 resistant isolate) and gentamicin MICs were 8–16 μg/mL (4/8 intermediate isolates and 4/8 resistant isolates). Molecular testing revealed that all the P. rettgeri isolates were positive for blaNDM-1 and blaPER-1. No amplicons were observed for the other primer pairs for blaVEB, blaCTX-M-1, blaCTX-M-8, blaCTX-M-9, blaSHV, blaTEM, blaKPC, blaVIM, blaIMP, blaOXA-10, and blaOXA-48. PFGE revealed that all isolates closely related one pulsotype with > 90% similarity (Fig. 2). The eight isolates had the three kinds of dendrogram patterns.

Table 2 Clinical characteristics of the outbreak cases and antimicrobial susceptibility profiles of Providencia rettgeri isolates
Fig. 1
figure 1

Time course of the outbreak by multidrug-resistant Providencia rettgeri. Black bars indicate the pre-infection period and gray bars the post-infection period in the surgical intensive care unit. Solid lines indicate the period during patients was hospitalized in a general ward

Fig. 2
figure 2

Pulsed-field gel electrophoresis patterns of Providencia rettgeri clinical isolate co-producing NDM-1 and PER-1. All eight isolates from the outbreak were closely related strains

Discussion

In the present study we reported and characterized an outbreak of blaNDM-1 and blaPER-1 carrying P. rettgeri. All patients were admitted to the same SICU and had a urinary catheter. P. rettgeri is well known to be isolated from urine of hospitalized and catheterized patients [16]. Although periods of hospitalization of our patients were not completely overlapping, PFGE revealed that all isolates were closely related. This suggests clonal cross-transmission of this strain in the SICU, and there is a possibility of transmission between patients and medical personnel by hand colonization or by environmental contamination. Infection control measures were reinforced in the SICU to include extensive environmental disinfection, active screening for carbapenemase-producing Enterobacteriaceae, and exhaustive contact isolation precautions. The outbreak did not eradicate in a short time, but the outbreak was eventually interrupted in January 2017.

Carbapenem resistance in Enterobacteriaceae has become a major public health challenge [28]. While carbapenem is a drug of choice for treatment of Enterobacteriaceae producing ESBL and plasmid-mediated AmpC cephalosporinase, production of carbapenemase in Enterobacteriaceae can be emerged. Carbapenemase gene is important due to its potential transferability to other species, by plasmids and transposons [28]. NDM-1 encoding plasmids are diverse and can also carry other antimicrobial resistance genes, including carbapenemase genes, ESBL genes, plasmid-mediated cephalosporinase genes, and aminoglycoside resistance genes [28, 29]. Among these, most ESBLs found with NDM-1 have been reported to be as CTX-M-15 type [29, 30]. Until now, this is the first report of Enterobacteriaceae co-carrying NDM-1 and PER-1 type ESBL. Although the NDM-1 enzyme is known to inactivate all β-lactams except aztreonam [6], our P. rettgeri isolates showed high MIC to aztreonam, possibly due to production of PER-1 type ESBL. The range of MIC to imipenem revealed 0.5–4 μg/mL. Imipenem MICs for Providencia spp. tend to be higher (e.g., MICs in the intermediate or resistant range) naturally. These isolates may have elevated imipenem MICs by mechanisms other than production of carbapenemases [18].

It is known that the multidrug-resistant bacteria have superior ability to survive and spread successfully in a hospital environment. In addition, the patient’s risk factor is also responsible for the nosocomial transmission of multidrug-resistant bacteria. Patient’s underlying disease, exposure to antimicrobial agents, and history of having invasive procedures are known as risk factors for the acquisition of carbapenem-resistant Enterobacteriaceae [28]. This outbreak persisted for 13 months, although the prompt infection control strategy was initiated after recognition of the first few cases. Because ICU admission patients often have one or more of risk factors, so it could be very difficult to eradicate once the outbreak occurs.

In conclusion, we report an alarming outbreak of high-level of multidrug-resistant P. rettgeri isolates co-producing NDM-1 and PER-1 β-lactamases. Infection prevention and control efforts should be continuously made to prevent nosocomial transmission of these threatening bacteria.