Distinguishing Clinical Enterococcus faecium Strains and Resistance to Vancomycin Using a Simple In-House Screening Test

Vancomycin-resistant enterococci (VRE) are a major concern as microorganisms with antimicrobial resistance and as a public health threat contributing significantly to morbidity, mortality, and socio-economic costs. Among VREs, vancomycin-resistant Enterococcus faecium (VREfm) is frequently isolated and is resistant to many antibiotics used to treat patients with hospital-acquired infection. Accurate and rapid detection of VREfm results in effective antimicrobial therapy, immediate patient isolation, dissemination control, and appropriate disinfection measures. An in-house VREfm screening broth was developed and compared to the broth microdilution method and multiplex polymerase chain reaction for the detection of 105 enterococci, including 81 VRE isolates (61 E. faecium, 5 E. faecalis, 10 E. gallinarum, and 5 E. casseliflavus). Verification of this screening broth on 61 VREfm, 20 other VRE, and 24 non-VRE revealed greater validity for VREfm detection. The accuracy of this broth was 100% in distinguishing E. faecium from other enterococcal species. Our test revealed 93.3% accuracy, 97.5% sensitivity, and 79.2% specificity compared with broth microdilution and PCR detecting van genes. The kappa statistic to test interrater reliability was 0.8, revealing substantial agreement for this screening test to the broth microdilution method. In addition, the in-house VREfm screening broth produced rapid positivity after at least 8 h of incubation. Application of this assay to screen VREfm should be useful in clinical laboratories and hospital infection control units.


Introduction
Enterococci are known as opportunistic pathogens and a leading cause of hospitalacquired infections, especially vancomycin-resistant enterococci (VRE) that are a major concern regarding their antimicrobial resistance as a public health issue. The two major species that are well-known worldwide to cause human diseases and resistance to vancomycin are Enterococcus faecium and E. faecalis [1]. Of these two species, vancomycin-resistant E. faecium (VREfm) is more frequently isolated from patients with hospital-acquired infection than vancomycin-resistant E. faecalis and other species [1].
The World Health Organization (WHO) published their list of priority bacterial pathogens for which new antibiotics are urgently needed, and VREfm is listed in the high-priority category of antibiotic-resistant bacteria [2]. VREfm is widely distributed in hospitals around the world, with the prevalence varying according to geographical location. In the U.S., VREfm accounts for 82% of reported cases because vancomycin is rarely restricted and thus has widespread antibiotic use in hospitals [3][4][5]. In European countries, VREfm has shown increasing prevalence, from 10% in 2015 to 17.3% in 2018 [3]. In Asia, the prevalence of VREfm accounts for 22.4% of reported cases and is higher than for European countries but lower than for the U.S. [6]. Treatment options for invasive VREfm infections are very limited, resulting in high mortality [7]. Vancomycin resistance determinants due to the vanA and vanB genes are globally frequently reported in VRE, including in E. faecium clinical isolates [3,8].
A previous report discussed potential determinants influencing the future dissemination and control of antibiotic resistance and nominated the rapidity and accuracy of laboratory techniques that allowed for rapid identification of the infecting pathogen and antibiotic susceptibility testing [9]. Therefore, accurate and rapid detection of VRE, especially VREfm, support effective antimicrobial therapy, immediate patient isolation, and appropriate disinfection measures in hospitals. Many methods have been implemented to identify either VRE or VREfm, including conventional culture, real-time PCR, conventional PCR, and automated microbiology instruments, such as BD Phoenix (Becton, Dickinson and Company) or Vitek-2 (bioMerieux) [10,11]. Although some of these methods are rapid, easy to perform, and are accurate in identification or detection, many of these methods are expensive and have a high level of sophistication that requires operator experience and skill and special instrumentation. These constraints make such methods inappropriate in a laboratory with limited resources. Therefore, we developed an alternative assay to detect VREfm isolates at a low cost. The aim of this study was to compare an in-house VREfm screening broth with the broth microdilution method and multiplex PCR for the detection of VREfm. Our screening broth provides an alternative assay to detect VREfm isolates. This may be useful in laboratories where a large number of isolates must be screened at low cost and as such could reduce the costs associated with laboratory and infection control in hospitals with a high prevalence of VRE, especially VREfm.

Results
In this study, we compared the in-house VREfm screening broth using the broth microdilution method and multiplex PCR for the detection of VREfm. All tested enterococci were confirmed to the species level, van genes detected, and the MIC values determined of vancomycin. The screening broth was evaluated using the broth microdilution method and PCR assay in terms of specificity, sensitivity, accuracy, and Cohen's kappa coefficient.
The vancomycin MIC values indicated that 85.9% (61/71) of the E. faecium isolates were resistant to vancomycin (range 32-128 µg/mL) that were confirmed to be VRE according to their MIC values and the presence of either vanA or vanB, whereas the remaining 10 isolates were susceptible to vancomycin. Of the 14 E. faecalis isolates, only five were resistant to vancomycin (VREfs) by broth microdilution, with 3 isolates having an MIC value of 64 or 128 µg/mL, and they all carried vanA, while the remaining two E. faecalis isolates carrying vanB revealed an MIC value of 32 µg/mL. In contrast with the 20 Enterococcus spp., only 10 E. gallinarum and 5 E. casseliflavus isolates showed either intermediate or full resistance to vancomycin with MIC values in the range 8-64 µg/mL, whereas the rest were susceptible (vancomycin-susceptible enterococci, VSE). The results are summarized in Table 1.

Evaluation of In-House VREfm Screening Broth
The in-house VREfm screening broth determined resistance to vancomycin based on turbidity and distinguished E. faecium from other enterococci based on color changes. The broth was used in two tubes (A and B). Tube A did not contain vancomycin, whereas tube B did (6 µg/mL vancomycin). Basically, all enterococci, including those that were vancomycin not-susceptible or susceptible, could grow in tube A, while growth in tube B depended on their resistance to the antibiotic. E. faecium could produce β-galactosidase, the enzyme degraded to Salmon-Gal that produced a red product, whereas the other Enterococcus spp. did not (Figure 2).
We found 5 VSEs (2 were VSEfm and 3 were VSEfs) that produced a positive result in tube B (containing vancomycin), indicating that they were vancomycin-resistant based on the screening broth. Indeed, they were susceptible to vancomycin based on broth microdilution with an MIC value of 4 µg/mL for all isolates (Table 1). In contrast, one isolate each of VIEg and VIEc showed neither turbidity nor color in tube B, although they showed intermediate resistance to vancomycin with an MIC value of 8 µg/mL (Table 1).
This suggested that our test had variation with the borderline MIC cut-off for some intermediate isolates, especially for VIEg or VIEc that carried vanC. However, there was good correlation of our screening broth and broth microdilution to indicate vancomycin resistance at MIC ≥16 µg/mL. The screening broth could identify E. faecium by producing the red color in all 71 isolates, while all the other species of Enterococcus (n = 34) were colorless. This demonstrated that the accuracy of this broth-based method was 100% for identification of E. faecium according to multiplex PCR assay. However, determination of vancomycin resistance and presence of van genes revealed 93.3% accuracy, 97.5% sensitivity, and 79.2% specificity (Tables 2 and 3). The kappa statistic to test interrater reliability showed 0.8 for this screening test. Based on the kappa criterion, our in-house VREfm screening broth had substantial agreement (values in the range 0.61-0.80) to the broth microdilution and PCR detecting van genes as the reference method. In addition, our in-house VREfm screening broth showed rapid positivity after at least 8 h of incubation, and the cost per test (2 tubes) was USD 0.9 or EUR 0.8.

Discussion
VREfm has a global impact as a threat to public health according to the WHO, and it contributes significantly to morbidity, mortality, and the socio-economic costs of healthcareacquired infection [2,12]. For example, the daily costs of contact isolation (considering only gloves and gowns) for each VREfm patient in a ward and an intensive care unit (ICU) were USD 10.8 and USD 17.3, respectively [13]. With the mean duration of isolation of 25 and 41.5 days for the ward and the ICU, respectively, a total cost was estimated to be USD 270 and USD 718, respectively [13]. Another study revealed that the overall hospital costs of blood-stream infection were significantly higher in VREfm cases (EUR 80,465) compared to VSEfm (EUR 51,365) and VSEfs (EUR 31,122) cases [14]. Screening is recommended for patients at high risk of VRE colonization to prevent and control the transmission of VRE. Our study successfully developed an in-house VREfm screening test to distinguish VREfm from other VRE and VSE isolates.
Validity testing based on the kappa coefficient demonstrated that our screening broth had strong agreement in the determination of VRE (especially VREfm) but false-positive or false-negative results may occur in either VIEg/VIEc or VSE isolates with low MIC values of about 8 or 4 µg/mL, respectively. For the false-positives in 5 VSE with MIC at 4 µg/mL, the actual MIC may be 6 or 7 µg/mL because the broth microdilution technique applied two-fold dilution of vancomycin that started from 0.25 µg/mL and produced 0.5, 1, 2, 4, 8, 16, 32, 64, and 128 µg/mL. The gap between 4 and 8 µg/mL may have contained the exact MIC for these 5 false-positive VSE isolates according to our screening broth, which contained 6 µg/mL of vancomycin, and so allowed these 5 isolates to grow. However, the false-negatives involving VIEg and VIEc had MIC values of 8 µg/mL (Table 2) and were PCR-positive for vanC (Table 3) but did not grow in tube B of the screening broth, even though we repeated both the broth microdilution and screening tests. The explanation for this has not been elucidated but it may depend on the strain characteristics or the low concentration of inoculum used that retarded bacterial growth under marginal conditions.
Another study evaluated chromogenic agar to screen VRE based on the different colors of colonies for either VREfm or VREfs [15]. A false-positive was detected in the medium with non-enterococci, such as Staphylococcus sciuri, Streptococcus mutans, or other enterococci (E. casseliflavus and E. raffinosus) [15]. Cross-reactivity of our screening broth may have occurred with S. sciuri, as mentioned in another study [15]. Performing catalase and Gram-stain testing of presumptive VRE isolates should increase the specificity. Notably, our screening broth identified E. faecium based on degradation of 6-chloro-3-indoxyl-β-Dgalactopyranoside by its β-galactosidase. This enzyme is also present in some species of enteroccocci, such as E. malodoratus, E. saccharolyticus, E. villorum, and E. dispar; however, these species have very low prevalence in hospitals, and their resistance to vancomycin is very rare [16,17]. To increase the efficiency and rapidity of VREfm screening among patients, the screening broth should be further studied and validated on clinical specimens directly. In addition, our screening broth has some limitations, including that it does not diagnose other pathogenic enterococci, such as E. faecalis, and does not specify the MIC. Furthermore, the broth could not be detected where there were strains resistant to teicoplanin only. Therefore, further development of this screening broth should be carried out to identify E. faecalis, E. gallinarum, or E. casseliflavus, to specify the MIC values, and to determine whether it is possible to detect teicoplanin-resistant strains.
The annual estimated laboratory costs were projected to be USD 19,074 for 6372 patients screened in a pediatric hospital between 2010 and 2014 in Turkey [13]. Therefore, screening priorities should be based on the prevalence of infection and the financial resources of the institution. The previous study revealed laboratory cost per specimen for rectal swab culture and PCR was USD 2.7 and USD 40.5, respectively [13]. The cost per test (2 tubes) for our assay was USD 0.9 or EUR 0.8, whereas the cost of the gold standard is about USD 2.7-3.0 or EUR 2.4-2.6. Application of this assay for VRE screening from pure culture should be useful for clinical laboratories and hospital infection control units where there is high prevalence of VRE to provide prompt information to facilitate the rapid control of VRE dissemination in hospitals and more rapid implementation of isolation precautions regarding VRE carriage or infection to other patients, as well as reducing the cost.

Bacterial Strains
In total, 105 enterococci, consisting of 71 E. faecium, 14 E. faecalis, and 20 other Enterococcus spp., were used in this study. These enterococci were isolated and sent by hospitals for further confirmation by the Public Health Microbiological Laboratory Service of the Faculty of Public Health, Kasetsart University Chalermphrakiat Sakon Nakhon province campus under the Emerging Antimicrobial Resistant Bacteria Surveillance Program (EARB). E. faecium ATCC BAA-2316 (vanA), E. faecalis ATCC51299 (vanB), E. gallinarum ATCC49608 (vanC1), and E. casseliflavus ATCC700668 (vanC2/C3) were used as controls in the multiplex PCR assay. E. faecalis ATCC29212 and E. faecium ATCC BAA-2316 were used as the control for the broth microdilution. In addition, E. faecium ATCC BAA-2316, E. faecium ATCC19434 and E. faecalis ATCC29212 were used for controls for the in-house VREfm screening broth.

Microbiological Analysis
Each isolate was cultured on sheep blood agar at 37 • C for 18 h and identified using a conventional biochemical test, including arabinose utilization, resistance to 6.5% NaCl, bile esculin degradation, and PYR (pyrrolidonyl β-naphthylamide) degradation [16], and multiplex PCR assay to simultaneously identify E. faecium and E. faecalis, as well as the vancomycin-resistant genes; vanA, vanB, vanC1, and vanC2/C3, that were developed in our laboratory [29,30]. Genomic DNA from all isolates was extracted using a NucleoSpin ® Tissue Kit (Macherey-Nagel, Germany) according to the manufacturer's instructions. The PCR reaction mixtures contained 1X JumpStart TM REDTaq ® ReadyMix TM Reaction Mix (Sigma-Aldrich, MO, USA) and 0.4 µM of each primer pair (IDT, Singapore; Table 4). The following PCR thermocycling parameters were used: initial activation of DNA polymerase at 95 • C for 3 min; 30 cycles of denaturation at 95 • C for 30 s, primer annealing at 55 • C for 30 s, and extension at 72 • C for 1.15 min; with a final extension at 72 • C for 5 min. The PCR products were resolved using gel electrophoresis for 30 min on 2% agarose gels (Vivantis, Selangor, Malaysia) in 0.5× TBE buffer (Vivantis, Malaysia). The gels were stained with ethidium bromide (Sigma-Aldrich, MO, USA) and visualized under ultraviolet light using a GeneGenius Bioimaging System (SynGene, Maryland, USA). The sizes of the PCR products were determined by comparison with a molecular size standard (GeneRuler™ 100 bp Plus DNA ladder, Thermo Fisher Scientific, Waltham, MA, USA) as shown in Figure 1. Susceptibility to vancomycin was performed to determine minimal inhibitory concentration (MIC) using broth microdilution according to the current CLSI guidelines [31]. A 0.5 McFarland standard suspension of the isolate was made from culture grown on Muller Hinton agar plates (BD BBL, Bergen County, NJ, USA). The interpretation of MIC followed the criteria in the CLSI 2021 guidelines, namely ≤4 mg/mL is susceptible, 8-16 mg/mL is intermediate, and ≥32 mg/mL indicates resistance [31]. Standard enterococci strains of E. faecalis ATCC 29212 (van absence) and E. faecium ATCC BAA-2316 (vanA presence) were used for quality control for the broth microdilution.

In-House VREfm Screening Broth
This screening broth was prepared using 2 tubes (A and B). Tube A (1 mL) contained 1.85% Brain Heart Infusion broth (Merck, Kenilworth, NJ, USA) and 0.2% 6-chloro-3indoxyl-β-D-galactopyranoside (Salmon-Gal; GoldBio, St. Louis, MO, USA). This concentration of Salmon-Gal was selected experimentally based on our optimization (data not shown). Tube B (1 mL) consisted of the same ingredients as tube A and also contained vancomycin (Sigma-Aldrich, St. Louis, MO, USA) at 6 µg/mL according to the 2021 CLSI guideline recommendation for VRE screening. Interpretation is shown in Figure 2. A 0.1-mL amount of inoculum at a concentration of 0.5 McFarland was inoculated into the broth.

Statistical Analysis
Diagnostic measures were calculated, such as sensitivity, specificity, and accuracy of each test. The kappa statistic was calculated to evaluate the associations and levels of agreement of the data [32].

Conclusions
We successfully developed an in-house VREfm screening broth to distinguish E. faecium from other Enterococcus spp., and to determine resistance to vancomycin in a single assay. This alternative screening procedure for VREfm could be useful for hospital infection control.

Data Availability Statement:
No new data were created or analyzed in this study. Data sharing is not applicable to this article.

Conflicts of Interest:
The authors declare no conflict of interest.