Case report

A 62-year-old male with non-ischemic cardiomyopathy requiring a left-ventricular assist device (LVAD) was hospitalized with LVAD driveline drainage. Computed tomography of the chest showed a phlegmon along the LVAD, including the outflow tract, pump, and driveline. On hospital day 6, he underwent surgical incision and drainage. Intra-operative cultures grew a vancomycin-susceptible (MIC 0.5 ug/mL by VITEK) Enterococcus faecium isolate. Intravenous vancomycin was started after debridement. On hospital day 15, he underwent repeat surgical incision and debridement with tissue flap reconstruction to achieve closure of his surgical wound. Additional intra-operative cultures obtained during this surgery grew E. faecium now with phenotypic vancomycin resistance (MIC >16 ug/mL by VITEK) with an otherwise identical susceptibility profile and detection of the vanA gene via molecular testing. VanA testing was not performed on the intra-operative swab specimen submitted from the first surgery because it did not meet laboratory criteria for testing, which in our center is isolation from blood or a sterile tissue/fluid specimen. Subsequent molecular testing of the initial phenotypically vancomycin-susceptible E. faecium isolate also revealed the presence of vanA, consistent with infection due to a vancomycin-variable enterococcal (VVE) isolate.

Mechanisms of vancomycin resistance

Vancomycin resistance is common among E. faecium isolates worldwide. ^{Reference O’Toole, Leong, Cumming and Van Hal2} At least nine different mechanisms of vancomycin resistance have been described in enterococci: vanA, vanB, vanC, vanD, vanE, vanG, vanL, vanM, and vanN. ^{Reference McEllistrem, Nordstrom, Lucas, Decker and Van Tyne3} Acquired vancomycin resistance is most commonly mediated by the vanA gene cluster through alteration of the glycopeptide binding site. The vanA gene cluster is found on transposon Tn1546 which is commonly incorporated into plasmids allowing for transfer and spread across strains and even species. ^{Reference Gagnon, Lévesque, Lefebvre, Bourgault, Labbé and Roger4} Other gene clusters (ie, vanB) similarly confer vancomycin resistance to enterococci and are named for the enzyme that ultimately modifies the glycopeptide binding site. These other gene clusters generally differ in the enzymes expressed by the operon, the modified terminal sequence at the glycopeptide binding site (eg, D-ala-D-lac vs. D-ala-D-ser), location on genetic elements, type of expression (inducible or constitutive), and degree of resistance conferred to different glycopeptides. The vanB gene cluster, for example, can confer resistance to vancomycin while retaining susceptibility to another glycopeptide, teicoplanin. The vanC gene cluster, found on chromosomes of E gallinarum and E. casseliflavus-E. flavescens, confers intrinsic low-level vancomycin resistance while retaining teicoplanin susceptibility. ^{Reference Courvalin5} Although other gene clusters have been described, they are much less common than vanA, vanB and vanC. Among E faecium isolates, vanA remains the most common mechanism conferring vancomycin resistance.

The vanA cluster contains several genes: van R, S, H, A, X, Y, and Z (Figure 1). The proteins encoded by these genes work together to change the glycopeptide binding site’s peptidoglycan terminal sequence (D-ala-D-ala) by replacing the terminal D-alanine with D-lactate, forming D-ala-D-lac. This alters vancomycin’s target site, inhibiting its mechanism of action. Each gene in the vanA cluster serves a different function. The protein complex VanR/S includes the regulatory (VanR) and sensory (VanS) components for the rest of the operon. VanR/S regulates vanHAX expression by forming a two-component signal transduction system. When VanS detects the presence of glycopeptides, it phosphorylates VanR to promote transcription of vanHAX. Enzymes encoded by vanHAX specifically alter the glycopeptide binding site. VanX cleaves the peptidoglycan terminal D-ala-D-ala, VanH is a dehydrogenase that produces lactate (D-lac), and VanA is a ligase that forms the bond between D-ala and D-lac. Although vanA, R, H, and X are considered essential for vancomycin resistance, vanY and vanZ have a less clear role. VanY is a D,D-carboxypeptidase that depletes the D-ala precursor thus favoring formation of D-ala-D-lac. VanZ performs an unknown function but may be more necessary for teicoplanin resistance than vancomycin resistance. ^{Reference Li, Walker and De Oliveira6}

Figure 1. Components of the vanA gene cluster and their function. Adapted from Faron et al. ^{Reference Faron, Ledeboer and Buchan30}

The vanB gene cluster is the second most common and clinically significant mechanism mediating vancomycin resistance in E faecium. The vanB gene cluster typically leads to vancomycin MIC values near the vancomycin breakpoint which may result in resistance. Like vanA, vanB gene products alter the peptidoglycan terminal sequence from D-ala-D-ala to D-ala-D-lac. ^{Reference Courvalin5,Reference Hashimoto, Kurushima and Nomura7} Although some of the genes are analogous to those of vanA, the vanB cluster primarily differs in its regulatory genes (vanR _B, vanS _B ). This difference is thought to allow vancomycin to induce expression of the vanB gene cluster, but not teicoplanin. ^{Reference Hashimoto, Kurushima and Nomura7}

Genotypic mechanisms associated with VVE and reversion to vancomycin resistance

A variety of mechanisms may lead to the genotypic-phenotypic mismatch found in VVE as well as the subsequent reversion to vancomycin resistance. First, several studies report frameshifts or complete or partial deletions of vanRS, inhibiting the organism’s ability to detect vancomycin and therefore promote vanHAX transcription. ^{Reference Gagnon, Lévesque, Lefebvre, Bourgault, Labbé and Roger4,Reference Szakacs, Kalan and McConnell8–Reference Sugumar, Peela, Viswanath, Walia and Sistla16} In these reports, transition from phenotypic vancomycin susceptibility to resistance was mediated by mutations that allowed vanH to utilize other promoter mechanisms instead of vanRS. In one example, vanHAX was incorporated into the chromosomal DNA of the organism allowing use of a constitutively active ribosomal RNA gene promoter. ^{Reference McInnes, Snaith and Dunn14} Others report mutations in the vanH promoter itself that allow for constitutive activity, obviating the need for activation by vanRS. ^{Reference Thaker, Kalan and Waglechner10} Second, another set of reports describe deletions in vanX leading to the genotypic-phenotypic mismatch in VVE. ^{Reference Hansen, Pedersen and Nielsen13,Reference Jung, Lee and Lee15,Reference Hammerum, Justesen and Pinholt17} These deletions result in an enzyme less proficient in cleaving D-ala-D-ala, thus allowing the isolate to retain susceptibility to vancomycin. Transition from phenotypic vancomycin susceptibility to resistance is mediated by alternative mechanisms to decrease the amount of D-ala-D-ala available, favoring formation of D-ala-D-lac. One such mechanism is an increase in vanA plasmid copy number to produce more VanX enzyme. Another includes mutations that inactivate ddl, an enterococcal ligase outside of the vanA operon that forms D-ala-D-ala. ^{Reference Gholizadeh, Prevost, Van Bambeke, Casadewall, Tulkens and Courvalin18} Finally, one report ^{Reference Sivertsen, Pedersen and Larssen1} describes silencing of vanHAX expression via an upstream insertion sequence, ISL3, as a mechanism to retain vancomycin susceptibility. Excision of ISL3 in the presence of vancomycin then accounted for the transition from phenotypic vancomycin susceptibility to resistance ^{Reference O’Toole, Leong, Cumming and Van Hal2} .

VVE has also been described amongst E. faecium isolates with vanB genotype. Hashimoto et al. reported single amino acid substitutions in the vanB cluster were associated with vancomycin susceptibility and demonstrated reversion to phenotypic vancomycin resistance upon vancomycin exposure. ^{Reference Hashimoto, Kurushima and Nomura7} The mechanism underlying reversion remains unclear but may be associated with increased transcription and thus expression of the vanB gene cluster and possibly mutations in the regulatory vanS _B gene ^{Reference Mowlaboccus, Shoby, Daley and Coombs19} or other mutations outside of the vanB cluster. ^{Reference Hashimoto, Kurushima and Nomura7}

Microbiology of VVE

Microbiology labs may use genotypic methods to detect vancomycin resistance among enterococcal isolates in addition to routine phenotypic antimicrobial susceptibility testing modalities. One common method is a genotypic assay for vanA in E. faecium, which has excellent sensitivity and specificity and can be performed more rapidly than conventional techniques. ^{Reference Tan, Jiang and Ng20} Genotypic testing for vanB is also commercially available. ^{Reference Zhou, Arends and Kampinga21} Both the Clinical and Laboratory Standards Institute (CLSI) ²² and the European Committee on Antimicrobial Susceptibility Testing (EUCAST) ²³ have acknowledged the potential benefits of genotypic testing, though neither organization has incorporated it into their recommendations for routine work. Because of this, individual laboratories implement different protocols detailing when and how to use genotypic testing in the work-up of enterococcal isolates.

Screening for vancomycin-resistant enterococci (VRE) may involve use of a vancomycin-impregnated chromogenic agar, which is a selective and differential media that uses color to distinguish E. faecalis and E. faecium from other enterococci with intrinsic vancomycin resistance. This method may be a preferred way of assessing VRE colonization in healthcare settings due to ease of throughput and low cost. ^{Reference Kling, Rios and Dirnberger24}

Early VVE detection is only possible when a combination of genotypic vanA testing and phenotypic susceptibility testing is used. Phenotypic methods such as disk-diffusion and traditional methods for determining a Minimum Inhibitory Concentration (MIC) will not detect VVE as these isolates are phenotypically susceptible to vancomycin. Studies have reported a range in vancomycin MIC from 0.38 to 2 μg/mL with MIC of the revertant commonly >256 (Table 1). Similarly, chromogenic agar will also not detect VVE. Neither the selective component of the chromogenic agar nor the dose and duration of vancomycin used during susceptibility testing induce the reversion mutations necessary to lead to a detectable vancomycin-resistant phenotype.

Table 1. Description of genetic mechanisms described for VanA-positive vancomycin-variable E. faecium isolates from published manuscripts from 2004 to 2023

NR, Not reported; MIC, Minimum Inhibitory Concentration.

Surveillance typically refers to surveillance for VRE using rectal cultures, with varying specific institutional protocols.

Clinical isolates were detected from clinical specimens including blood, urine, wounds, bile, and ascites.

CLSI ²² and EUCAST ²⁵ recommend that any Enterococcus isolate that tests positive for vanA or demonstrates resistance to vancomycin on routine antimicrobial susceptibility testing be reported as vancomycin-resistant. EUCAST additionally recommends reporting these isolates as teicoplanin resistant. This conservative approach accounts for the possibility of VVE as well as potential errors in conventional testing methods in the presence of discrepant results.

Epidemiology of VVE

One of the first well-described reports of VVE was from Quebec, Canada, where six cases were detected from rectal swab samples from hospitalized patients during routine VRE surveillance. ^{Reference Gagnon, Lévesque, Lefebvre, Bourgault, Labbé and Roger4} Of note, genotypic vanA and phenotypic susceptibility testing was routinely done on all isolates, allowing for the detection of VVE. Subsequent reports from Denmark ^{Reference Hammerum, Justesen and Pinholt17} and India ^{Reference Viswanath, Sugumar, Chandra Murthy Peela, Walia and Sistla26} confirmed that this organism is found worldwide (Table 1). Other reports describe the proportion of VVE among VRE isolates. Sazacs, et al. reported that 15% of patients from a 2-year VRE outbreak investigation in Toronto had VVE (44/285); 36% of these were identified during admission screening. ^{Reference Szakacs, Kalan and McConnell8} A subsequent study from a network of hospitals in Toronto reported that 47% (18/38) of vanA positive E.faecium clinical isolates were VVE. ^{Reference Kohler, Eshaghi and Kim27}

Like VRE, VVE has been associated with nosocomial outbreaks. Sivertsen, et al. reported an extensive outbreak investigation after two patients with VVE of the same sequence type (ST203) were identified in Norway. ^{Reference Sivertsen, Pedersen and Larssen1} Of 15,158 clinical and surveillance samples screened during an 18-month period, 93 were VVE and 1 clone dominated on pulse field gel electrophoresis, suggesting nosocomial spread. Sequencing of some of these isolates demonstrated a plasmid with a vanA gene cluster variant different from Tn1546 which is typically associated with vanA positive VRE. Importantly, the authors also detected this plasmid in one E. faecalis isolate, raising the possibility of horizontal inter-species transfer of mobile genetic elements. Hansen et al. reported another extensive outbreak investigation in Denmark in which the VVE clone (ST1421-CT1134) was identified and subsequently became the dominant vanA positive E. faecium clone in the country by 2019. ^{Reference Hansen, Pedersen and Nielsen13,Reference Hammerum, Justesen and Pinholt17} This same clone has been reported in outbreaks in Australia. ^{Reference Wagner, Janice and Schulz11}

Given the challenges of detecting VVE, it is likely that VVE from both screening and clinical samples is underreported. Available data also suggests that once this organism becomes endemic in an institution or region it is likely to spread and the routine microbiologic approach to evaluating vanA positive enterococci may need to be adjusted.

Clinical reports of VVE

Because most early clinical reports of VVE described treatment emergent vancomycin resistance, clinicians questioned if vancomycin was effective therapy for infections caused by these organisms. Coburn, et al. described a patient with vanA positive, vancomycin-susceptible E. faecium isolated from ascitic fluid who was found to have a clonal vancomycin-resistant E. faecium on rectal swab after 8 days of vancomycin therapy. ^{Reference Coburn, Low and Patel9} The two patients who prompted the outbreak investigation described by Sivertsen et al. both developed clinical failure and subsequently had vancomycin-resistant E. faecium isolated from wound or blood cultures after approximately 7 days of vancomycin therapy. ^{Reference Sivertsen, Pedersen and Larssen1} In a retrospective analysis comparing VVE to VRE and vancomycin-susceptible E. faecium cases, Kohler et al found no association between VVE and breakthrough bacteremia, but a majority of these cases were treated with agents active against VRE. They noted that acquisition of VRE and VVE had similar risk factors compared to patients infected with vancomycin-susceptible E. faecium. Those with VRE or VVE were more likely to have prior antibiotic exposure (though not exclusively vancomycin) and more likely to have a central venous catheter as a source of infection. Notably, these authors also found no association between VVE and 30-day mortality. ^{Reference Kohler, Eshaghi and Kim27}

Laboratory-based studies demonstrate reversion of VVE from a vancomycin-susceptible to a vancomycin-resistant phenotype, but it is unclear how frequently this reversion occurs within a bacterial population. ^{Reference Thaker, Kalan and Waglechner10,Reference Hammerum, Justesen and Pinholt17} Jung, et al. found that 22% (4/18) of VVE isolates reverted to vancomycin resistance upon laboratory exposure to either vancomycin or teicoplanin. ^{Reference Jung, Lee and Lee15} Wagner et al, found that after 48 hours vancomycin exposure, reversion to vancomycin resistance happened frequently, yet still below the threshold of detection for standard susceptibility testing. ^{Reference Wagner, Janice and Schulz11} These findings suggest vancomycin should be avoided when treating infections caused by vanA positive clinical isolates, regardless of phenotypic susceptibility testing, which is in line with current CLSI ²² and EUCAST ²⁵ guidelines.

Although phenotypic reversion has been described in the laboratory and clinical setting, further study is needed to determine which factors affect this phenomenon including site of infection and degree of vancomycin exposure (eg, duration and drug levels). Additionally, as many hospitalized patients receive vancomycin for other clinical reasons, we do not know how or if vancomycin exposure prior to clinical infection may further affect observed resistance patterns.