Corresponding author and reprint requests: P. C. Appelbaum, Department of Pathology, Hershey Medical Center, 500 University Drive, Hershey, PA 17033, USA E-mail: email@example.com
Methicillin-resistant Staphylococcus aureus (MRSA) is well-recognised as a major cause of infection in the healthcare setting but, even more worryingly, is now emerging in the community. The glycopeptides—notably vancomycin—have traditionally been the mainstay of treatment of MRSA but overuse has led to the emergence of vancomycin-intermediate and vancomycin-resistant MRSA (VISA and VRSA, respectively). Although the mechanisms underlying vancomycin resistance are not yet fully understood, changes to the bacterial cell wall—the site of action of the glycopeptides—are believed to be key. Recent evidence also supports the transfer of genetic material among bacteria as contributing to the development of VRSA. Based on the cases identified to date, risk factors for the development of VRSA may include older age, compromised blood flow to the lower limbs, and the presence of chronic ulcers. The true extent of the problem, however, remains to be determined—it is likely that many cases of VISA and VRSA infection go undetected because of suboptimal screening programmes and possible limitations of automated and non-automated detection methods. Effective screening directed at those patients considered to be most at risk should therefore be a priority. Not surprisingly, the spread of MRSA from the hospital to the community setting, coupled with the emergence of VISA and VRSA, has become a major cause of concern among clinicians and microbiologists. The treatment options available for these infections are now severely compromised and thus new classes of antimicrobial agents effective against MRSA, VISA and VRSA are urgently required.
Serious infections caused by Staphylococcus aureus are a worldwide phenomenon and occur in both the hospital and community settings. Initially, penicillin G was an effective therapy option for S. aureus infections. However, the emergence of penicillin resistance in S. aureus isolates over the past 50 years has had an impact on how such infections are treated . The development of methicillin resistance, in particular, has been a cause for concern among physicians and microbiologists in recent years and effective treatment options are diminishing. In the USA, most clinical S. aureus isolates are resistant to penicillin (95%) and over half are resistant to methicillin; the prevalence of methicillin-resistant S. aureus (MRSA) in Europe is also a problem [2,3]. The situation is complicated by the recent increase in serious infections caused by Panton–Valentine leukocidin toxin-producing community-acquired MRSA in the USA [4–7].
Until recently, the glycopeptide vancomycin represented a uniquely effective solution for treating infections caused by methicillin-resistant pathogens, including S. aureus. However, the overuse of this antibiotic in oral form for conditions such as pseudomembranous colitis has inevitably changed this situation. The first clinical isolate of vancomycin-intermediate-resistant S. aureus (VISA) was identified in 1997, and these strains have now been reported worldwide [8,9]. Although VISA strains have hitherto been thought to be rare, a recent study from Turkey, in which 46 of 256 (18%) MRSA isolates obtained (mainly from blood and pus) between 1998 and 2002 showed the VISA phenotype, suggests that their incidence may be on the rise . More recently, there have been reports of vancomycin-resistant S. aureus (VRSA), which is even more alarming, as these isolates demonstrate complete vancomycin resistance [11–14].
This review will briefly touch upon MRSA and its prevalence in both the hospital and community settings, before focusing on the emergence of VISA and VRSA onto the infectious disease scene. The mechanism of vancomycin non-susceptibility, the possible presence of resistant strains in healthy carriers, and laboratory methods for determining vancomycin susceptibility will also be discussed.
MRSA: a problem in both the hospital and community settings
MRSA is not only recognised as a major cause of infections within the healthcare setting but is also emerging as a cause of infections in non-healthcare settings (i.e., within the community). For example, MRSA clones were found in an outbreak of abscesses reported recently in a professional US football team . MRSA isolates from 9% of the team were compared with other community and hospital isolates. All were found to be carrying the gene for Panton–Valentine leukocidin and the gene complex for staphylococcal-cassette-chromosome mec (SCCmec) type IVa resistance, which are known to cause skin and soft tissue infections. Isolates were susceptible to most antimicrobial agents, but not β-lactams and macrolides . MRSA strains causing life-threatening bacteraemia in infants treated from birth in a neonatal intensive care unit have also demonstrated the genetic traits of community-associated MRSA . During 2003, 17 infants with bacteraemia due to S. aureus were tested for methicillin resistance: eight infants (47%) tested positive for MRSA, and isolates from six of these infants (75%) carried the SCCmec gene characteristic of community MRSA; four isolates were type IVa. All of the isolates carrying the SCCmec gene were resistant to β-lactam antibiotics and erythromycin, and one was also resistant to clindamycin. Seven (88%) of eight infants had septic shock, and vancomycin was used as initial therapy. Despite this, three infants (38%) died, and three had complications requiring prolonged antimicrobial therapy.
Worryingly, the above reports are likely to be just two examples of a more widespread clinical problem. Indeed, three MRSA pandemic clones have been traced to original isolates from Denmark, and dramatic increases in methicillin resistance among S. aureus strains, as well as increased numbers of MRSA infections, have been linked to the expanding reservoir of community-onset MRSA [16,17]. Furthermore, increased use of vancomycin to treat MRSA will result in increased vancomycin selective pressure in the community, which, in turn, may lead to more strains of VISA and VRSA.
Staphylococcal resistance: from VISA to VRSA
The concentration of vancomycin required to inhibit most strains of S. aureus is typically between 0.5 and 2 mg/L . S. aureus isolates for which vancomycin MICs are 8–16 mg/L are currently classified as vancomycin-intermediate, and isolates for which vancomycin MICs are ≥ 32 mg/L are classified as vancomycin-resistant . However, the Clinical and Laboratory Standards Institute (CLSI, formerly the National Committee for Clinical Laboratory Standards) recommended that S. aureus isolates with vancomycin MICs of 4 mg/L should be treated as susceptible. The Centers for Disease Control and Prevention (CDC) recommendations, on the other hand, are to consider these strains as potentially intermediate. They advocate retesting and investigation into the patient's history for vancomycin treatment and possible response to vancomycin therapy . There is also a question as to whether systemic infections (e.g., endocarditis) caused by S. aureus strains with vancomycin MICs of 2 mg/L are truly clinically vancomycin susceptible . Teicoplanin and vancomycin belong to the glycopeptide class of antibiotics and may therefore be expected to have a common resistance mechanism; the term glycopeptide-intermediate S. aureus is therefore synonymous with VISA .
Emergence of heterogeneous VISA and VISA
Over the last decade, S. aureus strains with elevated vancomycin MICs have emerged . In 1997, a Japanese group identified the first heterogeneous VISA strain—the precursor to VISA—with reduced susceptibility to vancomycin . The heterogeneous strain was isolated from the sputum of a 64-year-old man with MRSA pneumonia that was unresponsive after 12 days of vancomycin treatment. The MRSA (Mu3) strain was grown in a drug-free medium, and produced a sub-population of cells with varying degrees of vancomycin resistance, thus demonstrating natural heterogeneity, or variability, in susceptibility to vancomycin. Mu3 produced sub-clones in the presence of vancomycin, with resistance roughly proportional to the concentrations of vancomycin used. Selection of Mu3 sub-clones with 8 mg/L or more of vancomycin gave rise to sub-clones with a vancomycin MIC of 8 mg/L at a frequency of 10−6 CFU/mL or higher.
True VISA was also first reported in Japan in 1997, and the first isolate was obtained from a surgical wound in a 4-month-old infant who had undergone open-heart surgery . Two weeks after surgery, the patient became febrile and developed a purulent discharge from the surgical incision site. Despite vancomycin treatment for 29 days, the fever and discharge persisted. At this stage, an aminoglycoside (arbekacin) was added to the vancomycin treatment regimen for 12 days; the wound healed and antimicrobial therapy was discontinued. However, 12 days later the incision site appeared inflamed and an abscess developed. This was treated with arbekacin plus ampicillin–sulbactam, and debridement of the subcutaneous abscess was performed. The debridement sample revealed an MRSA strain (Mu50) with a vancomycin MIC of 8 mg/L by the broth microdilution method. This was the first clinical strain of S. aureus to demonstrate this level of vancomycin resistance and was quickly followed by reports of VISA emergence in other countries, including the USA, prompting a flurry of activity aimed at limiting its spread [9,20].
Emergence of VRSA
To date, four VRSA isolates have been identified in the USA: two from Michigan, and one each from Pennsylvania and New York [11–14]. The first clinical isolate was identified from a swab taken from the catheter exit site of a 40-year-old man in Michigan with diabetes, peripheral vascular disease and chronic renal failure in June 2002 . Prior to this, the patient had been treated for chronic foot ulceration, and had been the recipient of multiple courses of antimicrobial therapy, including vancomycin. He had developed MRSA following the amputation of a gangrenous toe in April 2002, and was treated with vancomycin and rifampicin. S. aureus strains resistant to vancomycin (MIC > 128 mg/L) and oxacillin (MIC > 16 mg/L) were isolated from the exit site, which appeared to be healed a week after the removal of the catheter, but the patient's chronic foot ulcer appeared to be infected. VRSA, vancomycin-resistant Enterococcus faecalis and Klebsiella oxytoca were isolated from the ulcer. This strain exhibited high resistance to vancomycin (MIC > 128 mg/L by broth microdilution). Shortly after this report, another clinical isolate of VRSA was described in Hershey, Pennsylvania, in September 2002 with an MIC of 32 mg/L by broth microdilution testing . The strain was isolated from a chronic foot ulcer in a 70-year-old morbidly obese hypertensive male patient who had received multiple courses of antibiotics (other than vancomycin) in the past and with a recent history of somnolence, intermittent fever, chills, malaise, night sweats and dyspnoea on exertion for several weeks before presentation . A third isolate of VRSA was obtained from a urine sample and the urinary tract catheter of an elderly patient in long-term care in New York in 2004 (MIC 32 mg/L by broth microdilution testing) . In addition, vancomycin-resistant enterococci were isolated from multiple sites in this patient (F. C. Tenover, personal communication). A fourth VRSA isolate has also recently been reported (March 2005), the second reported from Michigan . VRSA (MIC 256 mg/L) was isolated from a gangrenous toe wound in a 78-year-old male with a history of coronary artery disease, type 2 diabetes, peripheral vascular disease, neuropathy, chronic renal insufficiency and obstructive uropathy. Prior to the toe wound, the patient had received vancomycin for most of the 9 weeks following surgery for an aortic valve replacement procedure. Although a vancomycin-susceptible E. faecalis isolate was recovered from the toe wound prior to amputation, a vancomycin-resistant E. faecalis isolate was recovered from a surveillance rectal culture in the same patient. Thus, vancomycin-resistant enterococci have been isolated from three of the four VRSA patients, and in-vivo transfer of the VanA gene could have occurred in this setting. The pathogenesis of gene transfer in the Hershey patient is unknown.
A comparison of the first three of these VRSA strains to emerge is shown in Table 1. Factors that may be associated with VRSA infection are age, a compromised blood supply to the lower legs caused by conditions such as hypertension and diabetes, chronic foot ulcers and a history of prior vancomycin treatment.
Table 1. Comparison of three vancomycin-resistant Staphylococcus aureus (VRSA) strains
Glycopeptides (i.e., vancomycin and teicoplanin) exert their antimicrobial effects by inhibiting synthesis of the S. aureus cell wall . Cell wall thickening and, potentially, the transfer of genetic material are currently thought to underlie the development of vancomycin resistance. Vancomycin acts by binding irreversibly to the terminal D-alanyl-D-alanine of bacterial cell wall precursors, inhibiting cell wall production by attacking the sites responsible for cell wall synthesis .
Resistance in VISA strains is thought to occur as a result of changes in peptidoglycan synthesis. VISA strains synthesise extra peptidoglycan with increased quantities of D-alanyl-D-alanine residues. These residues bind vancomycin molecules and effectively sequester them, thereby preventing them from reaching their bacterial target (Fig. 1) [24,25]. Furthermore, the newly altered cell walls containing bound vancomycin further impede the progress of drug molecules.
A number of studies have investigated the role of the bacterial cell wall in vancomycin susceptibility in S. aureus[26–28]. Increased resistance to vancomycin in Mu50 S. aureus was associated with accelerated peptidoglycan synthesis, thickened cell walls, an increased proportion of glutamine non-amidated muropeptides and reduced peptidoglycan cross-linking . Thickening of cell walls correlated with the trapping of vancomycin in the outer layers and was considered to be the mechanism of resistance . The Mu3 isolates also displayed vancomycin resistance, with cell wall thickening and peptidoglycan synthesis activities, although neither of these occurred to the same extent as observed in Mu50 isolates . Other studies have suggested that structural and/or metabolic changes in cell wall teichoic acids may also play a role in the resistance mechanism by reducing the rate of cell wall degradation (instead of increasing the rate of cell wall synthesis), thus maintaining a correlation between wall thickness and decreasing susceptibility to vancomycin .
Recently, there has been evidence to support the exchange of genetic material among VRSA bacteria [20,25,30–32]. Genetic analyses suggest that the in-vivo transfer of vancomycin resistance from E. faecalis to an MRSA strain occurred to produce the Michigan VRSA isolate [30,31]. Acquisition of the vanA gene in the Michigan isolate occurred via the interspecies transfer of Tn1546 (the vanA transposon, harboured within a multiresistant conjugative plasmid) from co-isolated vancomycin-resistant E. faecalis. This isolate achieved vancomycin resistance by altering the terminal peptide of D-alanyl-D-alanine to D-alanyl-D-lactate, which only occurs with exposure to low concentrations of vancomycin, and the new dipeptide seems to have a reduced affinity for vancomycin (Fig. 2) .
Vancomycin resistance in healthy carriers
In a recent Brazilian study, vancomycin-resistant strains were detected among coagulase-negative staphylococci, both inside and outside the hospital environment . All strains (Staphylococcus capitis, Staphylococcus ureolyticus, Staphylococcus hemolyticus and Staphylococcus epidermidis) demonstrated unstable heteroresistance to vancomycin, and had significantly thicker cell walls than the control strains (p < 0.001). Unstable heteroresistance to vancomycin in these strains means that, if subcultured several times, they tend to revert to vancomycin susceptibility . However, in the presence of vancomycin, these revertant strains select for strains of vancomycin-resistant staphylococci at very high frequencies. This implies that, although these strains of vancomycin-resistant staphylococci may not disseminate with stable resistance, they can readily revert to vancomycin resistance when they are exposed to this antibiotic .
The presence of coagulase-negative staphylococcal isolates in healthy carriers is of concern, particularly outside the hospital environment, as there has been no systematic screening programme for staphylococcal vancomycin resistance in patients likely to be at the greatest risk (i.e., those in long-term care or nursing homes, or those with chronic leg or decubitus ulcers). Thus, we have no knowledge of how widespread this problem is until these urgently needed studies are undertaken. The consequence of reduced vancomycin susceptibility is clearly an increased possibility of antimicrobial failure. The high prevalence of MRSA and glycopeptide use, both thought to be risk factors for VRSA, make the widespread dissemination of these organisms an alarming and realistic possibility .
Testing for vancomycin susceptibility
Non-automated methods that are acceptable for detecting VISA and VRSA are: CLSI broth microdilution, agar dilution and Etest® with 0.5 McFarland standard to prepare inoculum (AB Biodisk, Piscataway, NJ, USA) . The Etest® has the advantage of showing small colonies around one or more zones of inhibition and is considered to be among the most discriminatory of these tests. Disk diffusion alone may be acceptable for VRSA but not for the detection of VISA . Fig. 3 shows disk diffusion and Etest® analysis of the Hershey VRSA isolate on Mueller–Hinton agar .
Automated methods [e.g., Vitek® and MicroScan® (bioMerieux, Hazelwood, MO, USA)] are not as effective, with two-thirds of confirmed VRSA isolates not being reliably detected by automated testing systems . Neither automated method accurately identified Hershey (Pennsylvania) and New York VRSAs with vancomycin MICs of 32–64 mg/L [11,14]. Thus, laboratories using these methods must also use a vancomycin agar screening plate test with brain–heart infusion containing 6 mg/L vancomycin .
Some S. aureus strains with vancomycin MICs of 4 mg/L, or even 2 mg/L, may not be truly clinically susceptible to vancomycin. Taking this into consideration, it may be timely for the CLSI and CDC to lower their recommendation for the concentration of vancomycin in screening plates. In Europe, the Société Française de Microbiologie (SFM) and the Deutsches Institut für Normung (DIN) recommend screening plates with 6 mg/L teicoplanin; this concentration is considered to be too high and a teicoplanin concentration of 1–2 mg/L would be more appropriate. The Hershey (Pennsylvania) VRSA isolate had a teicoplanin MIC of 4 mg/L .
MRSA has spread into the community and is now distributed across the world. Increased numbers of MRSA infections, plus greater methicillin resistance levels within MRSA isolates, have been linked to the expanding reservoir of these strains in the community. Increased use of vancomycin has led to selective pressure, and the subsequent appearance of VISA followed by VRSA has sparked alarm among physicians and microbiologists. Four isolated cases of VRSA infection have been reported to date and all of these have been in the USA. Based on this very small sample size, it appears that VRSA is more likely to affect older patients with a compromised blood supply to lower limbs and chronic ulcers. Directed screening of patients most at risk is urgently required to assess the extent of the VISA and VRSA problem as well as vancomycin non-susceptibility in coagulase-negative staphylococci. Caution must be exercised in the use of detection methods, as automated techniques alone are ineffective in detecting many of these isolates. With increasing antimicrobial resistance among bacteria, there is clearly a need for new classes of antibiotic with different mechanisms of action that are effective against MRSA, VISA and VRSA.