Resistance Mechanisms, Epidemiology, and Approaches to Screening for Vancomycin-Resistant Enterococcus in the Health Care Setting

Aminoglycoside resistance.Enterococcus spp. may demonstrate moderate-level (MIC, 62 to 500 μg/ml) or high-level (MIC, ≥2,000 μg/ml) resistance (HLR) to aminoglycosides (12). Moderate resistance is conferred by an intrinsic mechanism attributed to low permeability of the cell wall to the large aminoglycoside molecules (13). A study including 2,507 E. faecalis and 469 E. faecium isolates from clinical specimens in France found the prevalence of moderate resistance to be 8.9% and 49.2%, respectively (14). HLR to aminoglycosides is mediated through modification of the ribosomal attachment sites or the production of aminoglycoside-modifying enzymes (12). Both mechanisms are encoded by specific genes and are typically plasmid borne. The prevalence of HLR to aminoglycosides can vary from 40 to 68% among Enterococcus species isolates, depending on geographical location, and differs between E. faecalis and E. faecium (15, 16). Intrinsic resistance may be overcome using a combination treatment with a β-lactam and an aminoglycoside, also referred to as a synergy therapy approach. Synergy therapy is based on the principle that β-lactams will disrupt the cell wall, thereby facilitating increased cellular penetration by the aminoglycoside to reach a concentration sufficient for the inhibition of protein synthesis. This approach is effective in treating infections caused by organisms with moderate-level resistance to aminoglycosides; however, synergy is lost in strains with HLR to aminoglycosides, because the increase in intracellular aminoglycoside concentration cannot overcome the presence of specific aminoglycoside-modifying enzymes. Therefore, accurate differentiation between the moderate-level resistance (MLR) and HLR phenotypes is important when considering synergistic antibiotic therapies for serious infections (17). Laboratories can determine HLR by performing a high-level aminoglycoside test according to the CLSI guidelines.

β-Lactam resistance.Intrinsic resistance to β-lactams is maintained in enterococci through the overproduction of penicillin binding proteins (PBPs) with low binding affinity for β-lactams, most notably PBP 4 and PBP 5 (18–21). The level of resistance differs between E. faecalis and E. faecium, with E. faecalis being 10- to 100-fold less susceptible to penicillin than streptococci and E. faecium being at least 4- to 16-fold less susceptible than E. faecalis (22). Despite reduced susceptibility to penicillin, the majority of E. faecalis isolates (∼99%) remain susceptible to ampicillin, while <20% of E. faecium isolates demonstrate ampicillin susceptibility (23). Importantly, neither penicillin nor ampicillin susceptibility is indicative of susceptibility to cephalosporins. For patients who are aminoglycoside intolerant, combination treatment with ampicillin and ceftriaxone was observed to be as effective as ampicillin and gentamicin in treating VRE with the gentamicin HLR phenotype (24). These therapies are most useful in treating severe infections, such as endocarditis and meningitis (25).

Glycopeptide resistance.Glycoproteins are bactericidal drugs that function by binding to the terminal d-Ala-d-Ala in the pentapeptide portion of the N-acetylglucosamine (NAG)–N-acetylmuramic acid (NAM) peptidoglycan (PG) cell wall precursor (Fig. 1). Binding blocks transpeptide linkage of cell wall components, resulting in reduced integrity and, ultimately, cell death. Resistance to glycopeptides in Enterococcus spp. is mediated by the vancomycin resistance (Van) operon. This operon may be carried chromosomally or extrachromosomally on a plasmid. The Van operon consists of vanS-vanR, a response regulator; vanH, a d-lactate dehydrogenase gene, vanX, a d-Ala-d-Ala dipeptidase gene; and a variable ligase in which 9 variant genes have been identified (vanA, vanB, vanC, vanD, vanE, vanG, vanL, vanM, and vanN) (26). Expression is inducible by the two-component system VanS/R, which senses disruptions in the cellular membrane caused by glycopeptides, as well as cell wall damage caused by bacitracin or polymyxin B (27). The variable ligase gene is central in determining the level of vancomycin resistance (low, medium, or high), with the most commonly identified genes being vanA, vanB, and vanC. VanA is plasmid borne, confers high-level resistance (MIC, >256 μg/ml) to vancomycin, and is most commonly associated with E. faecium and E. faecalis, while chromosomally encoded VanC confers low-level resistance (MIC, 8 to 32 μg/ml) to vancomycin and is almost exclusively found in E. gallinarum, E. casseliflavus, and E. flavescens (11).

• Open in new tab
• Download powerpoint

FIG 1

Vancomycin acts by binding to d-Ala-d-Ala pentapeptides blocking cell wall biosynthesis. Resistance to vancomycin is conferred by the van operon, which consists of a two-component regulatory system (vanS-vanR) that responds to either vancomycin, disruption at the cell membrane, or both. Detection of stimulus then activates downstream genes vanH, vanA or vanB and vanX. VanX is a d,d-dipeptidase that cleaves d-Ala-d-Ala repeats, both depleting the pool of d-Ala-d-Ala and supplying the bacteria with a free d-Ala for Van(A/B). VanH is a d-hydroxyacid dehydrogenase that reduces pyruvate to d-Lac for the ligase Van(A/B). Van(A/B) then ligates d-Ala-d-Lac, allowing for the production of d-Ala-d-Lac pentapeptides that have low affinity for vancomycin. In addition, VanY is a d,d-carboxypeptidase that cleaves the d-Ala terminal peptide to further reduce pools of pentapeptides that have high affinity to vancomycin. Finally, VanZ, which is present on vanA-carrying strains, confers modest resistance to teicoplanin through an unknown mechanism. Differences in the level of resistance are likely a result of pentapeptide composition, as the ratio between pentapeptides consisting of high affinity to low affinity to vancomycin correlate with the isolate MICs. High-level resistance (HLR) occurs when pentapeptides are mostly composed of low-affinity molecules, and moderate-level resistance (MLR) involves more-heterogeneous pools of high- and low-affinity pentapeptides.

VanA resistance.Enterococcus spp. carrying vanA are highly resistant to glycopeptides and are the dominant VRE variants of E. faecium and E. faecalis globally. Resistance is mediated by substituting the high-affinity terminal d-Ala-d-Ala peptide on NAM subunits with d-Ala-d-Lac. This amino acid substitution causes a 1,000-fold decrease in the affinity of the pentapeptide for vancomycin (Fig. 1) (11). Incorporation of NAM subunits containing the substituted d-Ala-d-Lac peptide into the peptidoglycan layer requires PBPs other than PBP 4 and PBP 5. In the presence of vancomycin, these alternative PBPs become the dominant proteins for cell wall synthesis. The alternative PBPs used during the presence of vancomycin have enhanced binding to β-lactams, which when used together can allow synergistic treatment (28). VanA-expressing strains are also resistant to the glycopeptide teicoplanin (Te) due to the presence of an additional gene present on the vanA operon, vanZ, which confers resistance by an unknown mechanism (29).

VanB resistance.Isolates carrying vanB are less prevalent than vanA-carrying strains but can be found throughout the world and are commonly identified in Australia, where the majority of E. faecium VRE isolates carry vanB (30, 31). As with to vanA, resistance in vanB is mediated by converting d-Ala-d-Ala to d-Ala-d-Lac (11). However, vanB confers varied resistance to vancomycin, ranging from moderate- to high-level resistance (MIC range, 4 to >256 μg/ml) (11). The mechanism for the lower-level resistance compared to that of vanA is not well defined but is likely the result of a lower proportion of d-Ala-d-Lac substitution in the cell wall of strains carrying vanB. Resistance to vancomycin is proportional to the percent composition of d-Ala-d-Lac to d-Ala-d-Ala (32). Smaller amounts of d-Ala-d-Lac incorporation might result from reduced expression of the vanB operon, a reduction in VanX or VanB enzymatic activity, or a combination of minor mechanistic changes. Te resistance is not observed in vanB-carrying isolates, as vanZ is not encoded in this operon.

Risk of vancomycin resistance transmission to other pathogens.Vancomycin is one of the few antibiotics that can be used to treat infections resulting from Gram-positive multidrug-resistant organisms (MDRO), such as MRSA; therefore, transmission of vancomycin resistance from enterococci to MRSA is of major concern. Horizontal gene transfer has been shown to be a mechanism of transmission between enterococci, and in vitro studies have demonstrated that transfer between Enterococcus and S. aureus can occur (33). In 2002, the first case of vancomycin-resistant S. aureus (VRSA) was isolated from a 40-year-old diabetic patient undergoing dialysis and was obtained from a dialysis catheter cultured from both the exit site and catheter tip. A week later, the VRSA isolate was isolated from a diabetic foot ulcer. MIC testing and DNA sequence analysis conducted by the CDC confirmed the isolate as VRSA (34). To date, 14 cases of VRSA have been reported in the United States; however, the presumably frequent interaction between VRE and MRSA (cocultured, with the two organisms recovered from the same source) and rare incidence of VRSA isolation suggest that in vivo transfer of vancomycin resistance between these species occurs at an extremely low frequency (35). If a laboratory expects a potential VRSA isolate (vancomycin MIC, ≥16 μg/ml), local infection control at the hospital should be notified of the possible result, and the laboratory should work with local and state departments of public health to have the isolate tested for confirmation.

Horizontal gene transfer of the van operon between Enterococcus spp. and other organisms also appears to occur at a very low frequency. In a study that performed mating experiments between Enterococcus species and Gram-positive gut flora (Lactococcus spp. and Bifidobacterium spp.), no transfer of vanA between genera was observed (36). Interestingly, the authors observed that interspecies transmission within the Enterococcus genus, e.g., E. faecium to E. faecalis, occurred at a much lower frequency (1/10⁸ per donor/recipient) than intraspecies transmission (1/10⁶ per donor/recipient). This may explain the higher prevalence of vanA within E. faecium isolates, as transmission occurs between E. faecium isolates, but not between E. faecium and other Enterococcus species, at high frequencies. Infection control and antimicrobial stewardship programs that aim to reduce acquisition (i.e., colonization and/or infection) of VRE and MRSA infections may aid in reducing the possible transmission of glycopeptide resistance to S. aureus by minimizing coinfection with VRE and MRSA.

Epidemiology.Enterococcus spp. are normal colonizers of the human gastrointestinal (GI) tract but can be recovered from the skin, genitourinary (GU) tract, and the oral cavity. Surveillance studies conducted by the European Antimicrobial Resistance Surveillance System and National Healthcare Safety Network (NHSN) report varied rates of colonization, depending on geographical location and immune status of the patient population. Generally speaking, colonization rates for inpatients range from as low as <2% in Finland to as high as 34% in Ireland and 33% in intensive care unit (ICU) patients in the United States (37). Asymptomatic colonization is typically transient but can persist from months to years; decolonization is not recommended, since it can contribute to the accelerated development of resistance (38). Colonized patients are a potential risk to health care facilities, serving as a source of for transmission via shedding of VRE into the surrounding environment and onto health care workers (11, 39).

Hospital rooms housing patients infected or colonized with VRE are quickly contaminated and can also serve as a reservoir for transmission. Environmental sampling studies have isolated VRE from most surfaces found in rooms that previously housed infected patients, such as patient gowns, bedside rails, floors, doorknobs, and blood pressure cuffs (12). Enterococcus has been shown to persist for as long as 4 months on surfaces, and in a clinical setting, VRE was shown to persist through an average of 2.8 standard room cleanings, thereby acting as a continual source for possible transmission (40, 41). Recent studies involving environmental surveillance suggest that admission to a room previously occupied by a patient with VRE infection was associated with a significant increased risk for VRE acquisition by subsequent patients (hazard ratio, 4.4) (42).

Health care workers may also serve as vectors for the transmission of VRE between patients in different rooms or wards. Without proper hand hygiene, VRE can persist for up to 60 min on a health care worker (HCW)'s skin (11). Once hands are contaminated, hand hygiene plays a major role in reducing transmission rates; however, compliance is generally low. In a systematic review from 2010 to 2014 of several hand hygiene intervention studies, it was found that compliance ranged from 23 to 69% (43). Furthermore, it is important to educate HCWs about the possibility of contact with environmental services to ensure understanding that hand contamination can occur without specific contact with the patient, and precaution should be applied even in the absence of patient interaction.

Passive or active screening.Guidelines to control MDRO by the CDC recommend that all hospitals adopt active rather than passive screening approaches. In the case of VRE, passive screening refers to the detection of VRE from clinical specimens submitted for routine culture (i.e., not specifically submitted for detection of VRE). Patients are only isolated after laboratory detection (culture or molecular) or if previously positive for VRE on a prior admission. Because the ratio of VRE-infected to -colonized patients is 1:10, a passive screening approach may not adequately identify a sufficient proportion of colonized patients to effectively reduce VRE transmission (12). Modeling of screening strategies using data from a 10-bed ICU at the University of Maryland Medical Center estimated that passive screening would reduce VRE infection by only 4.2% compared with no screening or isolation practices (48).

Active screening is less well defined but can include the collection of specimens specifically for VRE screening or screening of stools collected for other purposes (e.g., for C. difficile testing). There are no specific recommended practices, but studies reporting on the utility of screening often include testing upon hospital admission for patients with specific risk factors (e.g., those who are immunosuppressed or have a history of broad-spectrum antibiotic usage) and patients received from long-term-care facilities. Additionally, continued periodic screening 1 to 2 times a week during hospital admission and a screen at discharge may be employed. In hospitals that do not perform specific screening for VRE colonization, reflex testing of C. difficile specimens for VRE can help determine the relative prevalence of VRE colonization in a patient population and can aid in early detection of outbreaks (49). Routine screening of patients at admission is effective in hospital settings with higher prevalence, but the cost of intensive screening may not be justified in hospitals with a low prevalence of VRE or in hospitals without high-risk patient populations. Implementation of active screening for all ICU patients at admission may reduce VRE infections by up to 39% compared with no screening (48). Another approach to reducing the spread of VRE within the hospital setting is to cohort patients by VRE colonization status, as determined by laboratory screening result, and dedicate specific HCWs to each cohort. This method may be difficult to coordinate, requires adequate hospital beds for cohorting, and is likely not practical for the majority of hospitals; however, reduced HCW contact between colonized and noncolonized patients has the potential to reduce the spread of VRE in the hospital (50).

Outcome studies of active screening.Estimating the real-world effect and benefits gained by implementation of routine screening strategies can be difficult, and clinical outcome studies are not without limitations. A lack of proper case controls or implementation of multiple variables at one time to provide optimal patient care may obscure the independent contribution of any single screening or infection control measure. In addition, results may be dependent upon site-specific factors, such as organism prevalence and compliance of HCWs, which vary from facility to facility. Further, infection control practices often differ between studies, and the materials and methods may not be adequate to generate optimal comparisons across studies. A summary of several studies that performed analyses of active screening is presented in Table 1.

Several studies have shown that active surveillance of stool or rectal swabs has a positive effect in reducing the burden of VRE and related infections in hospitals. A retrospective comparison between two urban hospitals in the United States demonstrated a 2.1-fold lower rate of VRE bacteremia in the hospital performing weekly active surveillance on all inpatients (51). Strain typing of all bloodstream infection (BSI) isolates revealed that 4 clones were responsible for more than 75% of all Enterococcus species BSI at the hospital that did not actively screen for VRE, compared to 37% at the hospital with active screening. These data suggest that hospital transmission may be significantly more frequent at hospitals that are unaware of patient colonization status. Similarly, implementation of routine screening of stool specimens coupled with hand hygiene and barrier protection for positive patients (in accordance with 2003 SHEA guidelines) reduced the rate of patient colonization by more than 80% (1.2% versus 0.17%) in just 3 years (52). A comparable reduction in VRE colonization was achieved by combining active screening and hand hygiene with tracking of previous carriers and environmental cleaning of colonized patients' rooms after discharge (53). Following implementation of this bundle strategy, the rate of VRE detection reduced from 1.5 to 0.5 per 1,000 admissions.

Counter to studies demonstrating measureable benefit, there are also reports in the literature that suggest that active screening of patients does not reduce the prevalence of VRE in a health care setting. One cluster randomized trial included multiple hospital ICUs across the United States. Screening for VRE was on all admissions, and ICUs were randomly assigned to one of two groups, (i) an intervention group (n = 10), which was given the result of screening and implemented barrier precautions for positive patients; or (ii) a control group (n = 8), which was blinded to the screening result and so no intervention was implemented (54). After adjustment for baseline characteristics, the incidence of colonization was not significantly different between the intervention and control groups (40.4 ± 3.3 and 35.6 ± 3.7, respectively, P = 0.35). Unfortunately, this study had several limiting factors, including a 3-day delay between the collection of specimens and reporting of results. This delay may reduce the impact of interventional practices, since VRE transmission from colonized patients may have occurred in the 3 days prior to the implementation of barrier precautions. In addition, adherence to barrier precautions was monitored and found to be suboptimal, with a median of 77 to 82% of HCWs using gloves or gowns, and only 69% compliance with hand hygiene after contact with colonized patients. Combined, these factors may have contributed to decreasing the potential impact of screening and infection control efforts. Importantly, this highlights the real-world difficulty in implementing an effective screening and infection control program and underscores the fact that laboratory screening alone is insufficient to significantly impact the rate of HAIs.

Cost analysis of active screening.Active screening for VRE can carry a substantial monetary burden for the laboratory, including the cost of materials and labor, that may be difficult to predict prior to implementation. The total cost of screening is dependent upon several factors, including the patient population that is screened (hospital wide versus select ICU populations), the method of screening (culture based versus molecular) and other factors, including the potential for automation. One laboratory supporting a 19-bed ICU performing rectal swabs on all patients at admission, weekly during hospital stay, and at discharge determined that the active screening cost to their laboratory was $30.66 per patient or $22,956 per year ($29,570 in 2015 using the Consumer Price Index), inclusive of labor, laboratory supplies for culture, specimen collection supplies, and processing costs (55). Implementation of additional infection control practices based on a positive screen result increases the overall cost to the hospital as well. In one study, this cost was estimated at $116,515 ($224/patient) annually for additional infection control measures in a 22-bed adult oncology unit (56). However, implementation of this program resulted in a net savings of $189,318 annually ($294,433 in 2015), based on a reduction in VRE BSI from 2.1/1,000 patient-days to 0.45/1,000 patient-days (estimated $123,081), reduced VRE colonization (estimated $2,755), and reductions in antimicrobial use (estimated $179,997). Compared with the previous reports, active screening alone is cost-effective if it reduces 1 to 2 VRE BSI events a year, whereas larger infection control programs may need to prevent multiple infections a year to be cost-effective. Long-term costs of specialized chromogenic VRE screening media can be offset if the laboratory adopts automated systems that can plate, incubate, and analyze media by reducing the cost of technologists' time (57).

Optimal specimen collection.When screening patients for a specific pathogen, it is essential that the optimal specimen is collected to achieve maximum sensitivity. The CDC recommends screening for VRE using either rectal swabs or stool specimens. The use of stool specimens submitted to the laboratory for C. difficile detection provides a noninvasive specimen and targets patients with factors that also predispose them to a higher VRE colonization rate. In one study, stools submitted for C. difficile testing had a VRE positivity rate of 10.4%, compared with 9.7% positivity for rectal swabs collected from high-risk (e.g., bone marrow and solid organ transplant and surgical ICU) patients based on culture using selective media (58). Other studies suggest that stool specimens may be superior to rectal swab specimens for recovery of VRE. A direct comparison of paired stool specimens and rectal swabs cultured using Enterococcosel agar (BD, Franklin Lakes, NJ) demonstrated a sensitivity of only 58% for rectal swab specimens compared to stool specimens (59). Enrollment for this study was low (n = 35 positive specimens), but additional dilution experiments found that rectal swabs collected from patients with ≤4.5 log₁₀ CFU of VRE/g of stool were unlikely to be detected by plating to selective media. Although potentially more sensitive, the clinical significance of these findings is unknown, since there have been no studies establishing the relative risk of transmission by patients harboring lower concentrations of VRE.

Culture-based screening methods.Laboratories that choose to offer screening for VRE need to consider test sensitivity, complexity, turnaround time, and cost. Chromogenic agars are relatively inexpensive and provide a valuable tool for the laboratory. These agars and have been successfully used to screen for MRSA colonization as a component of infection prevention efforts (60, 61). Several VRE chromogenic agars that can identify VRE from stool specimens are currently FDA cleared and commercially available. These media may also differentiate between E. faecium and E. faecalis. In general, chromogenic agars have superior sensitivity, ranging from 90 to 99%, compared to bile esculin azide agar containing vancomycin (BEAV), which is approximately 85% sensitive (62, 63). The specificity of chromogenic agars is also superior to BEAV, which can be as low as 70 to 75%. A summary of the performance of these agars can be found in Table 2 (62–68). False-positive results due to breakthrough growth of non-E. faecalis, non-E. faecium isolates are less common on chromogenic media than on BEAV, which frequently supports the growth E. casseliflavus, E. gallinarum, Leuconostoc spp., and Pediococcus species. These species harbor low-level vancomycin resistance and are also esculin positive, making them difficult to distinguish from E. faecalis and E. faecium. In a study comparing 5 chromogenic agars, non-VRE isolates were recovered from 16.5% of specimens using BEAV, whereas isolation of non-VRE organisms on chromogenic agar ranged from 0.9% to 5.6% of the specimens tested (62). These differences are likely due to the lower concentration of vancomycin used by BEAV, 6 μg/ml, versus 8 to 10 μg/ml used in chromogenic media, as well as the specific colony color imparted by the chromogens used. While more sensitive than BEAV, chromogenic agar still requires 18 to 24 h of incubation, which can delay the implementation of infection control measures.

Molecular screening methods.Nucleic acid amplification tests (NAATs) have been applied tor VRE screening and detection using both stool and rectal swab specimens. Among these assays are both laboratory-developed tests (LDTs) and commercially developed assays, such as the Roche LightCycler VRE (vanA and vanB) detection assay, the Cepheid Xpert vanA and vanB assay, and the Cepheid Xpert vanA assay. The Cepheid Xpert vanA assay is FDA cleared, and the Xpert vanA and vanB assay is CE marked. Currently, the Roche LightCycler VRE detection kit is marketed as research use only (RUO).

The performance of molecular assays is varied, with sensitivity ranging from 61.5 to 100% and specificity ranging from 14.7 to 99.5% (69, 70). The wide difference in assay performance is due to several factors, including the gold-standard method used for culture comparison and the prevalence of VRE in the study population. The Roche LightCycler VRE detection kit demonstrated a high negative predictive value (NPV) of 99.9% for VRE but suffered low positive predictive values (PPV) of 1.4% for vanB and 68.5% vanA when results were compared to direct culture using Enterococcosel agar (68). A significantly higher PPV of 92.0% was reported using the Cepheid Xpert vanA and vanB assay, but reference culture was performed using a broth enrichment step prior to plating, which may increase the recovery of VRE in culture (71). Interestingly, a similar study evaluating the Xpert vanA and vanB assay using broth-enriched culture reported a PPV of only 32% (72). These data suggest that the reference culture method can affect assay performance data, but additional factors also contribute to the poor PPV commonly observed with these assays.

A major factor impacting PPV and NPV calculations is the pretest probability or prevalence of a given target, in this case VRE, in the population being studied. In the two studies that reported a PPV greater than 90%, VRE prevalence was 30% and 47% among the specimens tested, whereas both studies reporting low PPVs of 1.4% and 2.6% had VRE prevalence of 1.3% and 1.4%, respectively (68, 69, 71, 73). These data suggest that the utility of molecular assays to serve as a “rule-in” test is limited; however, these assays can still be effectively used to identify noncolonized patients with a high NPV. An obvious advantage of molecular assays is the potential to reduce turnaround time (TAT) by 18 to 72 h compared with culture, thereby enabling rapid screening to rule out VRE. The short TAT may be especially beneficial for facilities that implement strategies where all new or returning high-risk patients are treated as VRE patients until a screen returns negative. However, the additional cost of primary screening using comparatively expensive molecular assays must be weighed against the cost of unnecessary isolation measures, which may include gowns, gloves, and private rooms.

Another factor contributing to the reported low specificity and PPV of NAATs is the presence of vanA and vanB in multiple bacterial species other than Enterococcus spp., including Streptococcus mitis, Streptococcus bovis, Eubacterium lenta, Ruminococcus spp., Lactococcus spp., Leuconostoc spp., and some Clostridium species (68, 74). Many of these organisms are ubiquitous colonizers of the human gastrointestinal (GI) tract, and vanB carriage can be high in both healthy and sick individuals. Graham et al. observed high rates of nonenterococcal vanB carriage in community adults (63%), children (27%), and hemodialysis patients (27%) (75). Currently available molecular tests do not link the van genes to Enterococcus; therefore, they cannot differentiate between vanA and vanB carried by Enterococcus versus other bacterial species. This shortcoming can be minimized through the use of an assay targeting only vanA, provided it is used in a geographic region with a low prevalence of Enterococcus spp. harboring vanB. Alternatively, preenrichment of specimens for 16 to 24 h in selective broth containing 16 mg/liter amoxicillin, 20 mg/liter amphotericin B, 20 mg/liter aztreonam, and 20 mg/liter colistin significantly increased Xpert vanA and vanB assay specificity when coupled with a reduced cycle threshold (C_T) for calling positive results (76). Following selective broth enrichment, a C_T of ≤25 demonstrated a sensitivity, specificity, PPV, and NPV of 97%, 100%, 100%, and 99.5%, respectively. This is compared to values of 100%, 77.3%, 41%, and 100%, respectively, for nonenriched samples using the assay default of C_T ≤35 for calling a positive result. Importantly, broth-enriched specimens with a C_T of >30 were reliably reported as true negative; however, specimens with C_T values of 25 to 30 were variable and required culture confirmation before reporting. In addition to requiring independent validation of modified specimen and C_T thresholds, broth enrichment also delays the reporting of results by 18 to 24 h. This delay abrogates one of the key advantages of molecular assays, namely, rapid TAT, and mirrors the time to result using chromogenic agars, which, in general, are less expensive per specimen than molecular assays. Finally, the use of NAAT as a primary screen for VRE means that organisms are not recovered and therefore are not available for strain typing to aid in outbreak epidemiology.