Molecular epidemiology of vancomycin-resistant Enterococcus faecium clinical isolates in a tertiary care hospital in southern Thailand: a retrospective study

Objective Vancomycin-resistant enterococci are nosocomial pathogens that are responsible for commonly causing healthcare-associated infections, and they exhibit increased resistance to many antimicrobials, particularly to vancomycin. The epidemiological data available on vancomycin-resistant enterococci (VRE) in Thailand are inadequate. Methods Using enterobacterial repetitive intergenic consensus-polymerase chain reaction (ERIC-PCR), this study investigated genes that encode antimicrobial resistance and genetic relatedness to further understand VRE prevalence. Ninety VRE isolates were collected between 2011 and 2019 from a tertiary care hospital in southern Thailand. Antimicrobial susceptibility was determined using the disk diffusion method and E-test methods. Multiplex PCR was performed to detect the van gene and virulence genes. Results The study showed a high prevalence of diverse multidrug-resistant VRE strains. The prevalence of VRE infection was the highest in 2014 (28 isolates, 39.4%). VRE were mostly found in the urogenital tract (26 isolates, 28.9%), followed by the digestive tract (20%), body fluid, i.e., pancreatic cyst fluid, peritoneal dialysis fluid, Jackson–Pratt (JP) drain (20%), and blood specimens (10%). Patients in medical and surgical wards had 71.1% multi-drug-resistant and 28.9% extensively drug-resistant (XDR) VRE strains, respectively. The most prevalent antibiotic resistance was to ampicillin (74.4%). Susceptibility to gentamicin and meropenem were similar (7% and 10%, respectively). Four isolates (4.4%) were resistant to colistin. Only vanA was detected among the strains. The virulence gene test showed that the detection rates of enterococcal surface protein (esp) and hyaluronidase (hyl) genes were 91.1% and 5.6%, respectively. According to ERIC-PCR analysis, 51 of 90 strains had clonality, with a similarity rate of 95%. Conclusions We conclude that there is a need to implement infection control practices and active surveillance. Molecular techniques can effectively detect antibiotic-resistant genes, which would allow monitoring to control VRE infection in hospitals.


Specimen collection and biochemical identification of enterococci species
Specimen types included blood, body fluid (e.g., pancreatic cyst fluid, peritoneal dialysis fluid, and Jackson-Pratt (JP) drain), urine, midstream urine, pus, and samples from the pelvic region, rectum, and tissue. Before use in experiments, all VRE isolates were preserved in 20% glycerol at −80 • C. Sampling of specimen was granted an exemption from requiring ethics approval by the ethical committee of the Faculty of Medicine, Prince of Songkhla University (REC-60-234-04-7).

Identification of Enterococcus faecium and E. faecalis at the species level by multiplex PCR
The genes encoding D-alanine ligases specific for E. faecium (ddl E. faecium) and E. faecalis (ddl E. faecalis) and 16S rRNA genes (rrs) were detected by a modified multiplex PCR technique using the primers in Table 2 (Dutka-Malen, Evers & Courvalin, 1995). The process used was initial denaturation at 95 • C for 10 min, denaturation at 95 • C for 45 s, annealing at 54 • C for 1 min, and extension at 72 • C for 1 min (35 cycles) (Bhatt et al., 2015). The PCR products were visualized on 1.0% agarose gel.

DNA extraction
Ninety isolates of VRE originating from specimens in a tertiary care teaching hospital (Songkhla, Thailand) were extracted for chromosomal DNA using a GF-1 bacterial DNA extraction kit (Vivantis Technologies Sdn. Bhd., Malaysia) in accordance with the manufacturer's instructions. DNA concentrations were quantified using a spectrophotometer at an absorbance of 260 nm (A 260 ). DNA purity was calculated from the ratio of A 260 and A 280 , and DNA quality was evaluated using agarose gel electrophoresis.

Detection of van genes and virulence genes
All isolates resistant to vancomycin were subjected to a multiplex PCR analysis to detect vancomycin-resistant and virulence genes using a specific primer as detailed in Table 2 (Bhatt et al., 2015;Bourgogne et al., 2007;Elsner et al., 2000;Vankerckhoven et al., 2004). PCR amplification was performed via initial denaturation for 3 min at 94 • C; 35 cycles of amplification consisting of 1 min at 94 • C, 1 min at 54 • C, and 1 min at 72 • C; and a final extension for 5 min at 72 • C. A 100-bp DNA ladder (GeneDireX, Germany) was used as a molecular-size marker.
ATCC 700221 (for vanA) and E. faecalis ATCC 51299 (for vanB) were used as positive controls. The PCR product was sent for sequencing (1st BASE DNA Sequencing Services, Selangor, Malaysia), and Basic Local Alignment Search Tool search was performed using the National Center for Biotechnology Information database.

Statistical analysis
Demographic data are presented as percentages (unless otherwise stated) and median values with interquartile ranges (IQR). All statistical analyses were performed using SPSS Statistics version 23 (IBM Corporation, Armonk, New York, US). Qualitative variables were compared using a t -test and Pearson's chi-squared test. A P-value <0.050 was considered statistically significant.

Ethical approval
This study protocol was approved by the Ethics Committee of the Faculty of Medicine, Prince of Songkhla University (Date: August 4, 2017; Approval ID: REC-60-234-04-7).

Characteristics of the study population and VRE sources
Of the 90 VRE isolates recovered from patients, 53 (58.9%) were from females and 37 (41.1%) were from males, for a female-to-male ratio of 3:2. Fifty percent of VRE positive patients were ≥65 years old, and the mean age was 64 years (IQR, 43.5-73; range, 3 months to 90 years). Among these isolates, 42 (46.7%) were obtained from the medical ward; 18 (20.0%) from the surgical ward; and 10 (11.1%) from the intensive care unit. Urine samples had the highest percentage of VRE of 28.9% (26), followed by those obtained from the rectal area at 20% (18) and body fluid at 20% (18). The period of collective sample and source of samples showed statistically significant difference with an antimicrobial pattern (P < 0.05), but no significant difference was observed in gender, age groups, and hospital units. The pooled prevalence of VRE infections was 9.6% (95% CI [3.948-16.052]). More epidemiological data are presented in Table 3.
On the basis of the diameter of the observed zone of inhibition, the result of antibiotic resistance showed that all the selected VRE isolates were found to be resistant to ampicillin, vancomycin, penicillin, tazocin, norfloxacin, and ciprofloxacin susceptibility. Of the 57 selected VRE isolates, 56 (98.2%) were resistant to imipenem. However, over 50% of selected VRE isolates exhibited resistance when tested for gentamicin, cefoperazone-sulbactam, ceftazidime, cefotaxime, ceftriaxone, colistin, and ertapenem (Table 4).
The determination of MIC levels for vancomycin to VRE isolates showed that all the strains of VRE were resistant to vancomycin with MIC ranging between 0.25 µg/ml and 256 µg/ml. Seventy-three isolates had MIC of 256 µg/ml, and seven had MIC of 128 µg/ml. One isolate was in the intermediate range with the MIC ranging between 8 µg/ml and 16 µg/ml (Table 5).

Detection of van determinants and virulence factor genes in VRE
Ninety VRE strains (100%) harbored the vanA gene, whereas the vanB, vanC, vanD, vane, and vanG genes were not found in any of these isolates. The esp and hyl genes were found in 82 (91.1%) and 5 (5.6%) isolates, respectively. However, gel, cyt, cpd, and ebp genes were absent from all isolates. The amplicons of the VRE isolates and resistance-gene distribution are shown in Fig. 1.

ERIC-PCR analysis
The ERIC-PCR patterns of the 90 VRE isolates are shown in Fig. 2    isolates; and I-N genotypes, two isolates. Thirty-nine isolates were uniquely genotyped. MDR and XDR isolates were distributed among all scattered patterns, except for the three genotypes of the XDR pattern (J, L, and N) ( Table 6). The heterogeneity among the isolates obtained from urinary tract infections was stronger than those from other infection sites.

DISCUSSION
Enterococci significantly contribute to nosocomial infections. These infections most often occur in health care settings, particularly in intensive care units (ICU). They can spread by person-to-person contact or from contaminated medical devices (Olawale, Fadiora & Taiwo, 2011). Control measures to prevent the spread of enterococci infections are essential. To monitor the spread of enterococci infections at our hospital, samples collected from patients are routinely screened by conventional biochemical tests for nosocomial enterococci (resistant or sensitive to vancomycin). The prevalence of VRE among the nosocomial enterococci isolates in our study was 0.2%-4.3%, resulting in an average prevalence of 1.2%. The prevalence of VRE isolates in our study was relatively low compared with that found in India (7.0%) (Bhatt et al., 2015) and in Egypt (6.3%) (Said & Abdelmegeed, 2019), Table 6   indicating that risks of VRE infections may be associated with geographic variability. Blanco et al. (2017) reported in 2017 that relative humidity and geographical location were factors affecting VRE colonization rates in adult patients admitted to ICUs in the United States. Increasing humidity resulted in a greater VRE colonization rate, and VRE colonization was lower in southern states. In developing countries, hospitals could inadvertently be increasing the risks of VRE colonization and infections because of the lack of oversight procedures for transferring patients from hospitals with VRE outbreaks (Kang et al., 2014;Resende et al., 2014). Risk factors for severity of the patient's condition, hemodialysis, and central venous catheter usage and in univariate analysis include male gender, alcohol use, and antibiotics. Hospitalization in certain units of the hospital and lack of isolation measures are primary factors that affect VRE acquisition (Fossi Djembi et al., 2017). Our study demonstrated similar results after our 90 patients, whose test samples were positive for VRE, were paired with demographic and clinical data from the patient database. Nearly 50% of VRE carrier patients were ≥65 years old, with a median age of 64 years, and were admitted to the medical ward. In a 2019 study by Mathis et al., older patients had the highest infection during VRE outbreak in 2013-2014 at a teaching hospital in Lyon, France (Mathis et al., 2019).

Code of isolates
In addition, positive VRE samples were detected in urine specimens (28.9%), followed by those obtained from body fluid (20.0%) and the rectal area (20.0%). These results were similar to a Brazilian study of VRE infection, where the highest VRE concentrations were isolated from patients with urinary tract infections (Resende et al., 2014). Vancomycinresistant E. faecalis is a significant cause of nosocomial infections stemming from the urinary tract, surgical wounds, and catheter use and affecting the bladder, prostate, and kidneys (Zalipour, Esfahani & Havaei, 2019).
Because VRE are multidrug-resistant organisms with limited therapeutic alternatives that result in increased morbidity and mortality in patients (Buetti et al., 2019;Zalipour, Esfahani & Havaei, 2019), we examined the 90 VRE isolates in our study for antimicrobial susceptibility. The findings showed resistance to β-lactam, glycopeptide, and fluoroquinolone antibiotics in most of the tested isolates, with the highest resistance rate to ampicillin (74.4%) and imipenem (62.2%).
Enterococci have intrinsic resistance to several classes of antibiotics that inhibit cell wall synthesis, such as β-lactams and glycopeptides, and they exhibit the ability to acquire new mechanisms of resistance, particularly to penicillin/ampicillin, aminoglycosides (highlevel resistance), and glycopeptides (Cetinkaya, Falk & Mayhall, 2000;Gagetti et al., 2019;Werner et al., 2008). Results also showed susceptibility to meropenem in 10 of 15 tested isolates and 7 of the 21 tested isolates were susceptible to gentamicin. Meropenem and gentamicin are β-lactam and aminoglycoside antibiotics, respectively, and combinations of meropenem and aminoglycosides could function as alternative therapies for multi-and VRE infections. Similarly, there is evidence that β-lactams in combination with daptomycin reduced bacterial activity (Smith et al., 2015). Moreover, this combination of antibiotic therapy was successful in the treatment of endocarditis/severe MDR enterococcal infection (Antony et al., 1997;El Rafei et al., 2018 Zaheer et al., 2020). In this study, all the detected VRE isolates were identified as E. faecium, a species rarely found in Songklanagarind Hospital. Other studies have found that vancomycin-resistant genetic elements can transfer between enterococcal species (Murray, 1998). Most VRE isolates were MDR, and only six VRE isolates were extensively drug-resistant (XDR). All VRE isolates in this study were vanA E. faecium strains, whereas vanB, C, D, and E were not detected, similar to the findings from a 2014 report from (Resende et al., 2014).
Enterococcal virulence that results in pathogenicity is considered a multifactorial process (Comerlato et al., 2013). Virulence factors promoting the emergence of hospital-acquired enterococcal infections have been reported in a number of studies. In our study, VRE strains originally isolated from urine specimen expressed a high frequency of the esp gene (91.1%), confirming the important role of esp in enterococcal colonization and biofilm formation in urinary tract infections (Comerlato et al., 2013). This finding was comparable to that of other investigators who reported that the incidence levels of esp in E. faecium isolates were 71.5% in northwest Iran (Sharifi et al., 2012) and 90% in China (Yang et al., 2015) but rather low in southwest Iran (57.5%) (Arshadi et al., 2018).
Other virulence factors, hyl, asa, gel, and cyt genes, were also analyzed. The hyl gene was detected in only five (5.6%) VRE isolates in our study, similar to the findings from Haghi et al. who found the hyl gene in 3 of 79 (3.8%) VRE isolates in northwest Iran (Haghi, Lohrasbi & Zeighami, 2019). We did not detect the asa, gel, and cyt genes, confirming that esp and hyl are primary virulence factors among clinical strains of vanA E. faecium and play an important role in pathogenesis.
Molecular techniques for typing bacteria can vary when comparing standardization, cost, reproducibility, discriminatory power, and interpretation. We found that ERIC-PCR is a straightforward procedure for distinguishing organisms that are closely related. This method is less expensive and faster than many other methods, such as PFGE and amplified fragment length polymorphism typing (Borgmann et al., 2007).
The ERIC-PCR results identified 14 clusters among the 51 VRE isolates. The majority of the isolates were clustered in the A genotype, mostly from urine samples, and collected from the medical ward, and these were resistant to ampicillin, gentamicin, and carbapenem, similar to a report from Columbia (Corredor et al., 2019). We identified 39 unique traits of VRE isolates in our study, implying a circulation of these strains in the Songklanagarind Hospital for a period of time prior to the present study. These findings may suggest a hospital epidemiology stemming from the spread of MDR and vancomycin-resistant E. faecium with endemic vanA VRE. These bacteria may contribute to antibiotic resistance and virulence through horizontal gene transfer.

CONCLUSIONS
This was the first study conducted in a tertiary care hospital in southern Thailand to show a high prevalence of MDR and XDR-VRE in patients, which could lead to pathogenic effects on public health. The results of this study suggest that overused antibiotics appear to be mediated mainly by the vanA gene, which carries one or two virulence genes that create a high resistance level to vancomycin. However, molecular techniques are highly effective in detecting antibiotic-resistant genes, which would allow monitoring to control VRE infection in hospitals.

ADDITIONAL INFORMATION AND DECLARATIONS Funding
This work was supported by the government budget of Prince of Songkla University, with a grant number of MED601339S. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.