VISA and hVISA in hospitals

Staphylococcus aureus is noted for its clinical spectrum of disease ranging from asymptomatic colonisation to overwhelming sepsis and death and for its ability to become resistant to antibiotics. Resistance to beta-lactams, methicillin resistance, was first described 50 years ago, becoming a clinical problem in hospitals in the 1970s and the community in the 1990s. MRSA strains that originated in hospitals are usually also resistant to most of the non-beta-lactams as well, leaving vancomycin as the main parenteral drug to treat serious MRSA infections, with the role of new drugs like daptomycin and linezolid not well defined. MRSA strains can exhibit low-level resistance to vancomycin (vancomycin-intermediate S. aureus [VISA]), probably due to a thickened cell wall, which results in the trapping of vancomycin away from the active site of the septum in dividing cells. Detecting this resistance is difficult as multiple genetic pathways lead to this resistance, obviating a molecular test, forcing reliance on phenotypic tests, all of which have issues with sensitivity, specificity and cost. Mortality of bloodstream infection correlates with vancomycin MIC so in this situation the MIC should be determined by Etest or microbroth dilution especially if endocarditis is present. Detection of resistant subpopulations (heterogeneous vancomycin-intermediate S. aureus [hVISA]) can be done with the expensive and time-consuming population analysis profile (PAP) but it is unclear if this confers additional therapeutic information.

There are two essential types of reduced vancomycin susceptibility in S. aureus: (1) that conferred by cell wall thickening resulting in a small rise in the MIC and/or the presence of subpopulations with a modestly elevated MIC¹; and (2) that conferred by horizontal gene transfer of the vanA gene complex from vancomycin-resistant enterococci (VRE) to S. aureus, causing a more marked rise in the vancomycin MIC and clinical failure².

The Clinical and Laboratory Standards Institute (CLSI) reduced the breakpoints for vancomycin resistance in 2006 in response to the evidence that modest elevations of MIC were associated with a reduced likelihood of clinical response, and currently define vancomycin-susceptible S. aureus (VSSA) strains as having a MIC ≤4 mg/L, VISA as having MICs 4–8mg/L, and vancomycin-resistant S. aureus (VRSA) strains as having an MIC ≥16¹. The European Committee on Antimicrobial Susceptibility Testing (EUCAST) defines VSSA as strains with an MIC ≤2 and VRSA as having an MIC ≥4, with no intermediate division¹.

The first described clinical isolate of VISA was the so-called ‘Mu50’ strain, isolated from a child with an MRSA sternotomy infection that failed to respond to vancomycin but responded to debridement and arbekacin plus ampicillin/sulbactam; the microbroth vancomycin MIC was 8 mg/L³.

Mu50 and subsequent VISA isolates exhibit common characteristics (Figure 1)¹. They grow more slowly than VSSA, exhibit pleomorphic colonial morphology, have reduced or delayed coagulase, demonstrate thickened cell walls on electron microscopy, have alterations in cell wall metabolism, reduced susceptibility to lysostaphin and decreased autolysis.

Figure 1. hVISA strains exhibit colonial pleomorphism, in this case on a horse blood agar plate (a). Electron microscopy of vancomycin-susceptible Staphylococcus aureus (b) and heterogeneous vancomycin-intermediate Staphylococcus aureus (c) strains show how the cell wall is thickened in the latter.

hVISA strains usually have a vancomycin MIC in the CLSI and EUCAST susceptible ranges but possess subpopulations with a raised MIC, detectable by performing a PAP, where serial 10-fold dilutions of a suspension of S. aureus are plated on a series of agar plates containing increasing dilutions of vancomycin¹. This was first described by Hiramatsu’s group on the so-called ‘Mu3’ strain⁴, and Wootton et al. refined this by using a series of plates with doubling concentrations of vancomycin and serial 10-fold dilutions of the inocula with Mu3 as the control, generating the categories of VSSA (PAP-AUC ratio test: Mu3 <0.9), hVISA (ratio 0.9–1.35) and VISA (ratio >1.35) (Figure 2)⁵.

Figure 2. Population analysis profiling is currently the gold standard test for heterogeneous vancomycin-intermediate Staphylococcus aureus (hVISA) in the absence of a genetic test or phenotypic test with high enough sensitivity and specificity. The method generally requires a large number of plates (a), generates several plates requiring colony counts (b), and then entry of the numbers into a spreadsheet to generate a PAP graph (c), calculate the area under the curves for the hVISA control (Mu3) and the test isolate. In this case the test isolate, our first hVISA strain, gave a curve almost indistinguishable from Mu3.

Just over a dozen patients to date have had MRSA isolated that possess the vanA gene complex². Most are US origin; none have been reported from Australasia to date. The US patients are older with multiple co-morbidities, were colonised with vanA VRE and had high inoculum infections at sites difficult to penetrate with systemically administered vancomycin to which they have been extensively exposed. The vanA genes appear to have moved via plasmids from VRE into MRSA, conferring high-level resistance to vancomycin, therapeutic failure and major infection control issues.

Hiramatsu has proposed the affinity-trapping hypothesis to explain vancomycin resistance in VISA/hVISA strains⁶. VISA/hVISA strains increase the cell wall thickness and vancomycin this binds to the outer layers of the cell wall and it is not available to penetrate into the deeper layers and thus interfere with cell division.

VISA/hVISA strains do not possess any elements of the vanA, vanB,vanC, etc., gene complexes found in enterococci that confer vancomycin resistance¹. Many isolates have mutations in the walKR operon, which results in low activity, reduced autolysis and an increase in cell wall thickening¹. The mprF/fmtC mutations result in reduced cell wall cross-linking¹. There are also mutations in the graR genes and possible vraR as well, which results in upregulation of the cell wall stimulon¹.

Mwangi et al. applied whole genome sequencing to five sequential MRSA human blood culture isolates and found 35 point mutations across the five isolates, including genes involved in the cell envelope stress response, mutations in the agr quorum sensing system, the walKR wall operon and genes involved in peptidoglycan biosynthesis⁷. The group also detected mprFyycG, rpoB and rpoC gene mutations that were associated with daptomycin resistance (the patient was naïve to this drug). Cameron et al., performed comparative genomics in a seven sets of vancomycin exposed isogenic S. aureus pairs and found that serine/threonine phosphatase stp1 contributes to reduce vancomycin susceptibility to vancomycin, in addition to previously described mutations such as vraG, agaR, yvqF and rpoB⁸.

A recent systematic review and meta-analysis of the clinical significance of the vancomycin MIC in MRSA bloodstream infections compared mortality where the MIC was ≥1.5 µg/mL to isolates where the MIC was <1.5 µg/mL and found an odds ratio of death of 1.64 (95% CI 1.14–2.37) in the higher MIC group⁹.

Holmes et al. compared patients with MRSA bloodstream infection where the vancomycin MIC was 2.0 or 3.0 versus MIC of ≤1.5 and found an all-cause 30-day mortality of 27% versus 12.5% (P < 0.001). The association with vancomycin MIC also occurred within the groups treated with flucloxacillin and in those with methicillin-susceptible S. aureus; the authors suggested that the raised vancomycin MIC might be an epiphenomenon¹⁰.

Currently the clinical significance of hVISA status is unclear. van Hal and Paterson performed a systematic review and meta-analysis of the clinical significance of hVISA¹¹. They found hVISA infections had a clinical failure rate 2.37 times that of those with VSSA (95% CI 1.53–3.67), but all-cause 30-day mortality in the two groups was not significantly different. Peleg et al. described the greater wax moth (Galleria mellonella) model in which S. aureus strains with reduced susceptibility to vancomycin were associated with decreased pathogenicity and this correlated with decreased agr functional status¹².

Testing can be done using a number of methods, as summarised in Tables 1 and 2.

Table 1. Methods to detect VISA (based on van Hal and Fowler¹³).

Table 2. Methods to detect hVISA (based on van Hal and Fowler¹³).

Figure 3. The macromethod Etest (MET) heterogeneous vancomycin-intermediate Staphylococcus aureus (hVISA) screen uses a 2.0 McFarland inoculum and 48 hours of incubation to select resistant subpopulations. A positive is defined as a reading for teicoplanin ≥12, or if the teicoplanin MIC is ≥8 and the vancomyin reading is also ≥8. Note that these numbers are not the MIC. In this case, the teicoplanin (a) and vancomycin (b) readings are both ≥8; this is the strain whose PAP is shown in Figure 2c, where it was confirmed as a hVISA. Note also the appearance of small colonies close to the Etest strip, and the morphological variation in the colonies, which is also typical of hVISA strains.

Howden et al. in a comprehensive review outlined possible approaches to detection of VISA/hVISA¹. If laboratory testing for VISA/hVISA is readily available, they suggest a laboratory-based approach, in which a vancomycin broth MIC could be used or alternatively modified Etest or an Etest GRD with confirmation of positives with a PAP, and if reduced susceptibility to vancomycin was demonstrated the treating doctor should consider daptomycin or linezolid. If VISA/hVISA testing is not readily available, a clinical approach could be adopted where a tardy clinical response might prompt the use of alternative agents such as daptomycin or linezolid. In clinical practice in the treatment of MRSA bacteraemia clinical failure often prompts the use of these agents before results of VISA/hVISA testing are available.

van Hal et al. compared different testing methodologies for detection of hVISA¹⁴. Four hundred and fifty-eight consecutive MRSA blood culture isolates were tested with PAP, macromethod Etest, glycopeptide resistance detection, GRD Etest and vancomycin MICs by Etest, by broth microdilution using CLSI criteria and Vitek2. Four hundred and fifty-eight isolates from 470 episodes of blood stream infection were analysed; 55 were hVISA and four were VISA by PAP, the latter 4 were excluded. The sensitivity and specificity of the various methods compared with PAP were Etest 91% and 66%, broth microdilution 89% and 84%, modified Etest 89 and 55%, broth microdilution (cutoff ≥2 mg/L) 82% and 97%, Etest (cutoff ≥2 mg/L) 71% and 94%, GRD Etest 71% and 94%, and the Vitek2 (MIC estimate ≥2 mg/L) 25% and 96%, respectively. PAP, modified Etest, GRDE test, broth microdilution, standard Etest and Vitek2 cost $70–$320, $29, $12, $8–$38, $15 and $18 per isolate, respectively. Etest tended to read higher and Vitek2 lower than BMD. Performing PAP on all the isolates would have 100% sensitivity and specificity (assuming PAP is the gold standard) but cost $32,000 to test all MRSA isolates in the study. Screening with Etest with cutoff MIC ≥1.5 confirming positives with PAP gave 98.9% sensitivity and cost ~$20,000. MET followed by PAP gave a sensitivity of 89% but cost ~$30,000. Other test combinations were cheaper but insensitive.

Holmes et al. recently summarised key points about the clinical and laboratory implications of the different types of vancomycin-reduced susceptibility¹⁵. For methicillin-susceptible S. aureus infections, β-lactams give better results than glycopeptides. For invasive MRSA infections, optimal vancomycin dosing and source control are pivotal. In this situation an MIC should be determined, preferably by BMD but Etest was satisfactory. If there was clinical failure of vancomycin and/or one of the recommended methods suggested reduced vancomycin susceptibility, daptomycin or linezolid should be considered. If the MIC ≥16 mg/L polymerase chain reaction testing for vanA and vanB genes should be performed as the isolate may be a VRSA.

van Hal and Fowler summarised the studies looking at MIC creep, which refers to the slow increase in the MIC of the MRSA isolates over period of time¹³. Creep is mostly demonstrated if Etest is used and not found in most studies using broth microdilution. Studies that demonstrated MIC creep tended to be single-centre studies. Typing was only performed in a small number of studies and where this was done, creep was demonstrated where there was a clonal emergence of strains with increased MICs, which probably explains the phenomenon.

Daptomycin is a large molecular weight compound with a site of action in the cell wall, like vancomycin, and it has been hypothesised that VISA/hVISA strains might be less susceptible to daptomycin. A patient treated unsuccessfully with vancomycin but never treated with daptomycin, has a series of MRSA isolates from blood demonstrating increasing resistance to vancomycin and also daptomycin¹⁶. Sakoulas et al. found MRSA strains that were heterogeneously resistant to vancomycin were also heterogeneously resistant to daptomycin¹⁷. Another group found MRSA MICs to daptomycin and vancomycin were strongly correlated (P > 0001, χ² test)¹⁸.

If there is clinical failure of glycopeptide treatment of MRSA infection, especially bacteraemia, generally daptomycin or linezolid will be selected before screening or definitive testing for VISA/hVISA can be performed. In the case of MRSA bacteraemia, especially if endocarditis or other deep focus is present, the MIC should be determined, either by Etest (not modified Etest which is a screen for hVISA) or broth microdilution.

Fortunately most MRSA infections do not involve bacteraemia. Vancomycin is cheap, generally well tolerated, has defined therapeutic drug monitoring strategies, and works well for the parenteral treatment of most MRSA infections.

The clinical significance of hVISA needs to be defined. A cheap reliable test that predicts clinical failure with vancomycin is needed. The role of daptomycin, linezolid and the new anti-MRSA cephalosporin ceftaroline in serious MRSA infections needs to be clarified.