Heterogeneous Vancomycin-Intermediate Susceptibility Phenotype in Bloodstream Methicillin-Resistant Staphylococcus aureus Isolates from an International Cohort of Patients with Infective Endocarditis: Prevalence, Genotype, and Clinical Significance

Abstract

BackgroundThe significance of heterogeneous vancomycin-intermediate Staphylococcus aureus (hVISA) is unknown. Using a multinational collection of isolates from methicillin-resistant S. aureus (MRSA) infective endocarditis (IE), we characterized patients with IE with and without hVISA, and we genotyped the infecting strains

MethodsMRSA bloodstream isolates from 65 patients with definite IE from 8 countries underwent polymerase chain reaction (PCR) for 31 virulence genes, pulsed-field gel electrophoresis, and multilocus sequence typing. hVISA was defined using population analysis profiling

ResultsNineteen (29.2%) of 65 MRSA IE isolates exhibited the hVISA phenotype by population analysis profiling. Isolates from Oceania and Europe were more likely to exhibit the hVISA phenotype than isolates from the United States (77.8% and 35.0% vs 13.9%; P<.001). The prevalence of hVISA was higher among isolates with a vancomycin minimum inhibitory concentration of 2 mg/L (P=.026). hVISA-infected patients were more likely to have persistent bacteremia (68.4% vs 37.0%; P=.029) and heart failure (47.4% vs 19.6%; P=.033). Mortality did not differ between hVISA- and non-hVISA–infected patients (42.1% vs 34.8%, P=.586). hVISA and non-hVISA isolates were genotypically similar

ConclusionsIn these analyses, the hVISA phenotype occurred in more than one-quarter of MRSA IE isolates, was associated with certain IE complications, and varied in frequency by geographic region

Infections caused by heterogeneous vancomycin-intermediate Staphylococcus aureus (hVISA) are poorly understood [1]. Defined as the presence of subpopulations (typically at a rate of 1 organism per 10⁵–10⁶) of methicillin-resistant S. aureus (MRSA) with intermediate vancomycin resistance [2], the hVISA phenotype has now been reported in clinical MRSA isolates from many parts of the world [3–7]. However, fundamental questions persist involving hVISA, including its prevalence, global distribution, and clinical significance

Nowhere is the issue of antibiotic efficacy more important than in infective endocarditis (IE) caused by MRSA. Vancomycin remains the first-line treatment for this condition, often out of necessity rather than choice. However, its efficacy is hindered by its slow bactericidal mechanism of action and poor penetration into valvular vegetations [8, 9]. Furthermore, several studies have reported an association between the hVISA phenotype and deep-seated endovascular infections, such as IE [3, 7, 10, 11]. Because of these observations, we reasoned that patients with MRSA IE would be at particularly high risk for the presence of the hVISA phenotype and that bloodstream isolates from such patients would be ideal to address many of the unresolved issues regarding the prevalence, global distribution, and clinical significance of hVISA. However, the relative infrequency of MRSA IE in any single center made a study focusing on this infection impractical until now

The current study had 2 major objectives: (1) to define the prevalence and clinical significance of the hVISA phenotype in MRSA IE and (2) to evaluate the genotypic characteristics of MRSA IE bloodstream isolates exhibiting the hVISA phenotype. To address these issues, we used MRSA bloodstream isolates obtained from a large, contemporary, international cohort of prospectively identified and clinically well-characterized patients with definite IE

Patients and Methods

Patients and settingsData from the International Collaboration on Endocarditis–Prospective Cohort Study (ICE-PCS) were used for this study. The ICE-PCS Microbiological Repository contains >1500 bloodstream isolates of bacteria obtained from prospectively identified patients with definite IE from >16 countries between 13 November 1999 and 20 January 2006 [12, 13]. For the current study, bloodstream isolates from all patients with definite MRSA IE were eligible for inclusion

DefinitionsDefinite IE was defined according to modified Duke criteria [14]. Other terms were defined as follows: chronic immunosuppressive therapy, administration of recognized immunosuppressive agents (including oral corticosteroids or other agents, such as those used in solid organ transplantation or rheumatologic disorders) for >30 days at the time of IE diagnosis; cardiac device, permanent pacemaker, cardioverter-defibrillator, or prosthetic cardiac valve; intravascular access device, arteriovenous fistula or indwelling vascular catheter; chronic indwelling central catheter, tunneled or cuffed catheter, or subcutaneous port catheter; stroke acute neurological deficit of vascular etiology lasting >24 h [15]; vascular or immunologic evidence of endocarditis, physical examination findings of endocarditis present at initial evaluation, including the Osler nodes, Janeway lesions, Roth spots, conjunctival hemorrhage, splinter hemorrhages, or peripheral vascular embolic events; congestive heart failure, criteria in the New York Heart Association classification system [16]; persistent bacteremia, >3 days of bacteremia despite receipt of an antibiotic to which the isolate was susceptible in vitro; health care–associated IE, nosocomial or nonnosocomial health care–associated infection. Nosocomial infection was defined as IE developing in a patient hospitalized for >48 h before the onset of signs or symptoms consistent with IE, and nonnosocomial health care–associated infection was defined as IE diagnosed within 48 h of admission in an outpatient with extensive health care contact, that is, one who had (1) received intravenous therapy, wound care, or specialized nursing care at home within the 30 days before the onset of S aureus IE; (2) attended a hospital or hemodialysis clinic or received intravenous chemotherapy within those 30 days; (3) was hospitalized in an acute care hospital for ⩾2 days in the 90 days before the onset of S. aureus IE; or (4) resided in a nursing home or long-term care facility [17]. The remaining clinical, echocardiographic, and outcome variables were defined as reported elsewhere [12]

Laboratory methods and susceptibility definitionAll S. aureus isolates were identified by means of standard methods. The minimum inhibitory concentration (MIC) of vancomycin was determined by Etest (AB Biodisk) according to the manufacturer’s instructions and by the broth microdilution method according to the manufacturer’s protocol (Siemens) and to Clinical and Laboratory Standards Institute recommendations [18]. Except where noted, all vancomycin MIC values in the text were defined according to Etest. Reference strain American Type Culture Collection (ATCC) 29213 was used as a positive hVISA control strain. The MIC of daptomycin was determined by Etest according to the manufacturer’s instructions. All MICs were interpreted by a single blinded investigator (J.F.F.)

Population analysis profilingAll MRSA isolates were assessed by the population analysis profiling–area under the curve (PAP-AUC) method, using the technique described by Wootton et al [19]. Organisms were cultured in tryptone soy agar plate from an overnight growth, adjusted to a density of 10⁸ colony-forming units/mL, and spiral-plated (Eddy Jet Spiral Plater) onto brain-heart infusion agar (Difco) plates containing 0, 0.5, 1, 1.5, 2, 3, 4, and 8 mg/L vancomycin. Colonies were counted after a 48-h incubation at 37°C, and the viable count was plotted against the vancomycin concentration with Prism software (version 5.0; GraphPad). This plot was then used to calculate an area under the curve. Isolates were defined as hVISA if they had a PAP-AUC ratio of >0.9 and <1.3 compared with the Mu3 reference strain (ATCC 700698)

Multiplex polymerase chain reactionGenomic DNA was prepared as described elsewhere [20]. Bacterial determinants, including adhesins, toxins, agr groups I–IV, and other genes, were screened by multiplex polymerase chain reaction (PCR), as described elsewhere [20, 21]. The isolates were classified as staphylococcal chromosomal cassette mec (SCCmec) type I, II, III, or IV by multiplex PCR, as described elsewhere [22]

Pulsed-field gel electrophoresisPulsed-field gel electrophoresis (PFGE) with SmaI was performed on all isolates, and the gels were analyzed using BioNumerics software (version 4.0; Applied Maths), as described elsewhere [23]

Multilocus sequence typingMultilocus sequence typing (MLST) was performed as described elsewhere [24, 25]. PCR fragments of 7 housekeeping genes (arcC, aroE, glpF, gmk, pta, tpi, and yqiL) were obtained from chromosomal DNA and were directly sequenced. MLST allele names and sequence types were derived from the MLST database (http://www.mlst.net). Clonal complexes were assigned to groups of isolates sharing 6 of 7 alleles by means of eBURST (http://eburst.mlst.net)

StatisticsSimple descriptive statistics were used to describe the demographics and other characteristics of the patients and to provide the genetic and antibiotic sensitivity profile of the bacterial isolates. Statistics are presented as medians with interquartile ranges for continuous variables and as frequency counts and percentages for categorical factors. The statistical significance of associations between variables was calculated using the Kruskal-Wallis test for continuous measures and the Fisher exact test for cross-classifications of categorical data. For all tests, differences were considered significant at P<.05. All statistical analyses were performed with SAS software (version 9.1.2; SAS Institute)

Results

Patient characteristics and outcomeIsolates from a total of 65 prospectively enrolled patients from 13 centers in 8 countries with definite MRSA IE were available for this study. Of these, 19 (29.2%) exhibited the hVISA phenotype by the PAP-AUC method. Patients with IE due to hVISA were older (median age, 73.1 vs 60.6 years; P=.037) (Table 1), more likely to be from Oceania or Europe (P<.001) (Figure 1), and more likely to have native valve IE (94.4% vs 53.3%; P=.005). None of the 65 patients were injection drug users. Patients with hVISA had a higher rate of persistent bacteremia (68.4% vs 37.0%; P=.029) and congestive heart failure (47.4% vs 19.6%; P=.033). Rates of other complications and in-hospital mortality did not differ significantly between groups

Table 1

Clinical Characteristics of and Outcomes in 65 Patients with Infective Endocarditis Due to Heterogeneous Vancomycin-Intermediate Staphylococcus aureus (hVISA) or Vancomycin-Susceptible S. aureus (VSSA)

Figure 1

Prevalence of the heterogeneous vancomycin-intermediate Staphylococcus aureus (hVISA) phenotype among methicillin-resistant S. aureus bloodstream isolates from patients with infective endocarditis, by geographic region. Percentages of patients with hVISA phenotype are specified above each column

Patient treatmentRates of glycopeptide therapy did not differ between groups; such therapy was used in 43 (93.5%) of 46 patients infected with vancomycin-susceptible S. aureus (VSSA) and 18 (94.7%) of 19 infected with hVISA (P=1.00). Of the 43 VSSA-infected patients treated with glycopeptides, 2 received teicoplanin. All 18 hVISA-infected patients treated with glycopeptides received vancomycin. Twenty-two (51.2%) of 43 VSSA-infected patients treated with a glycopeptide also received combination antimicrobial therapy: vancomycin plus rifampin (7 patients), vancomycin plus rifampin and aminoglycoside (6 patients), vancomycin plus aminoglycoside (3 patients), vancomycin plus trimethoprim-sulfamethoxazole (3 patients), vancomycin plus linezolid (1 patient), teicoplanin plus quinupristin-dalfopristin (1 patient), or fosfomycin plus aminoglycoside (1 patient). Eleven (61.1%) of 18 hVISA-infected patients receiving vancomycin also received rifampin (5 patients), aminoglycoside (3 patients), fosfomycin plus imipenem (2 patients), or linezolid (1 patient). No statistically significant differences between VSSA- and hVISA-infected patients were noted in rates of surgical therapy (30.4% vs 15.8%; P=.353) or mean time to surgical treatment (20.5 vs 6 days; P=.130, Kruskal-Wallis test). The frequency of patients transferred from another facility also did not differ between VSSA- and hVISA-infected patients (20 [43.5%] of 46 vs 8 [42.1%] of 19; P>.99)

Genotypic characteristics of hVISA isolatesSeveral genes were highly conserved in both hVISA and VSSA isolates (Table 2Table 2⁴). For example, spa, ebps, hlg, and efb were present in all of the tested isolates. The hVISA isolates were significantly more likely to contain the genes for fnbA (47.4% vs 15.2%; P=.011) and seh (21.1% vs 2.1%; P=.023) than were the VSSA isolates. The frequency of agr type II polymorphism did not differ between the hVISA and VSSA isolates (P=.586). The distribution of SCCmec types of hVISA isolates was significantly different from that of VSSA isolates (P=.012), with SCCmec type III being more common in hVISA isolates (21.1% vs 0% for VSSA; P=.006)

Table 2

Genotypic Characteristics of Heterogeneous Vancomycin-Intermediate Staphylococcus aureus (hVISA) and Vancomycin-Susceptible S. aureus (VSSA) Isolates from 65 Patients with Methicillin-Resistant S. aureus Infective Endocarditis

Table 4

Characteristics of and Outcomes in 65 Patients with Methicillin-Resistant Staphylococcus aureus Infective Endocarditis, According to Vancomycin Etest Minimum Inhibitory Concentration (MIC) of the Infecting Isolate

By MLST, all but 4 of the 65 isolates were included within 1 of 4 clonal complexes or sequence types: clonal complex (CC) 5 (31 isolates [47.7%]), CC8 (11 isolates [16.9%]), CC30 (11 isolates [16.9%]), and sequence type 228 (7 isolates [10.8%]). The distribution of these groups did not differ considerably between the hVISA and non-hVISA groups. Of the 48 MRSA isolates with a typeable PFGE profile by USA typing schema, 26 isolates (54.2%) demonstrated a USA100 pattern, and 10 isolates (20.8%) showed a USA200 pattern (Figure 2). Three isolates demonstrated the USA300 genotype; none exhibited the hVISA phenotype. The distribution of PFGE profiles among the hVISA isolates and VSSA isolates did not differ significantly

Figure 2

Dendrogram of pulsed-field gel electrophoresis (PFGE) profiles of 65 methicillin-resistant Staphylococcus aureus bloodstream isolates from patients with infective endocarditis. The clonal complex (CC) or sequence type (defined by multilocus sequence typing), geographic region of origin (region), and heterogeneous vancomycin-intermediate S. aureus (hVISA) vancomycin-susceptible S. aureus (VSSA) status of each isolate is also represented

hVISA, vancomycin MIC, and clinical outcomeThe hVISA phenotype was significantly associated with higher vancomycin MIC values as determined by Etest (P=.026), with all but 1 of the hVISA isolates having a vancomycin MIC of ⩾1.0 mg/L. In addition, the hVISA phenotype was present in 5 of the 6 isolates (83%) with a vancomycin MIC of 2 mg/L (Table 3). To confirm this association between the vancomycin MIC and the hVISA phenotype, we repeated the analyses using MIC values as defined by the broth microdilution method. These repeated analyses again indicated a significant association between increasing vancomycin MIC and the presence of hVISA phenotype (P=.006, Fisher exact test) (Table 3). All 65 MRSA isolates were susceptible to daptomycin, with a daptomycin MIC of ⩽1.0 mg/L by Etest (data not shown). We also evaluated potential associations between clinical outcome and higher vancomycin MIC values. As shown in Table 4, no significant differences in outcome were identified among patients with IE due to MRSA isolates with varying levels of vancomycin MICs

Table 3

Percentage of Heterogeneous Vancomycin-Intermediate Staphylococcus aureus (hVISA) Isolates by Population Analysis Profile, According to Vancomycin Minimum Inhibitory Concentration (MIC) as Defined by Etest and Broth Microdilution

Discussion

More than one-quarter of the isolates in this study exhibited the hVISA phenotype. This rate— ∼3 times higher than those reported in 2 recent reports of patients with primarily non-IE infections [6, 26]—may reflect the fact that IE is a high-inoculum infection, with valvular vegetations serving as “sanctuary sites” for the bacteria. Penetration of glycopeptides into these vegetations is limited [27], leading to the potential for prolonged bacterial exposure to subtherapeutic glycopeptide levels. The presence of hVISA related to prolonged subtherapeutic glycopeptide exposure seems particularly possible among MRSA isolates with higher vancomycin MICs [28]. In support of this possibility is the observation that >80% of the IE isolates in the present investigation with a vancomycin MIC of 2 mg/L exhibited the hVISA phenotype

hVISA-infected patients had more IE complications, including persistent bacteremia and congestive heart failure, than did patients with VSSA IE. Congestive heart failure is a leading cause of death in patients with IE [29, 30], and its presence has the greatest impact on prognosis of all IE complications [31, 32]. The association between persistent bacteremia and hVISA has been previously reported among patients with S. aureus infections of multiple types [7, 10, 11] but was confounded by the issue of whether the persistent bacteremia was simply due to the presence of IE. The current study addressed this issue by demonstrating that—even among patients with IE—persistent bacteremia and hVISA are associated. On the other hand, it should be acknowledged that persistent bacteremia does not equal failure per se, particularly because the mortality difference between VSSA- and hVISA-infected subjects (35% vs 42%) did not achieve statistical significance. Because this observation is consistent with other reports [7, 10], it is possible that some infections due to hVISA may eventually be treated successfully with vancomycin. However, the recent finding by Rose et al [33] that daily vancomycin doses of up to 10 g were ineffective against an in vitro model of a high-inoculum hVISA infection argues against this possibility. Thus, our failure to demonstrate a significant difference in mortality between hVISA- and VSSA-infected patients may simply be due to an insufficient sample size. However, the fact that almost 40% of the entire cohort died underscores the need for alternative therapies for MRSA IE

The prevalence of hVISA among MRSA IE isolates differed by geographic region in our study. IE isolates from Oceania or Europe were more likely to exhibit the hVISA phenotype than were isolates from North America. Although the absolute number of isolates and study sites from each region was relatively small, the findings were striking: more than three-quarters of MRSA IE isolates from Australia or New Zealand exhibited the hVISA phenotype by the PAP method. Previous studies evaluating hVISA prevalence have used a variety of laboratory definitions, testing strategies, and patient populations [6, 10, 34–39] (Table 5), thereby limiting comparisons. The current investigation overcomes these limitations to make a key point: even in a “high-risk” infection type such as IE, the prevalence of hVISA varies significantly and is potentially associated with nonclinical factors, such as geography. Although the reasons for the apparently higher hVISA prevalence in Oceania and Europe are unknown, it is intriguing to note that teicoplanin (a glycopeptide that has been associated with development of the hVISA phenotype [40, 41]) is used in both of these regions but not in the United States. There may be a connection, but further studies will be required to substantiate this interesting speculation

Table 5

Prevalence of Heterogeneous Vancomycin-Intermediate Staphylococcus aureus (hVISA) in Other Studies

Detection of the hVISA phenotype in S. aureus is difficult. The PAP-AUC method is time-consuming, labor intensive, and unsuitable for clinical laboratories. Several other laboratory methods have been considered [42–44] but also have limitations as simple screening tests. Thus, a clinically directed approach will likely be required before hVISA testing can be routinely considered. One readily available strain characteristic that may identify bacteria at higher risk for hVISA is the presence of a vancomycin MIC of 2.0 mg/L. Although only ∼10% of the cohort had a vancomycin MIC of 2.0 mg/L by Etest, >80% of these isolates exhibited the hVISA phenotype by PAP testing. Thus, discovery of a MRSA isolate with a high vancomycin MIC from an appropriate clinical setting (eg, MRSA IE with persistent bacteremia) could identify situations in which hVISA testing might be helpful. However, any such screening strategy should be properly validated before widespread adoption in clinical practice

Using PFGE, MLST, and PCR to identify a large number of putative virulence genes, we found that hVISA isolates were genotypically similar to VSSA isolates. This finding conflicts with another report suggesting an association between agr type II polymorphism and the development of vancomycin heteroresistance [45]. Our failure to identify a single clonotype associated with hVISA also suggests that the emergence of hVISA may be due more to the influence of ubiquitous forces (eg, widespread glycopeptide use) acting simultaneously on multiple different bacterial clones in different regions of the world than to the global dissemination of a limited number of bacterial clones exhibiting hVISA. This finding also suggests that bacterial genotype is unlikely to be a useful method to target selected clinical isolates for hVISA testing. However, a recent study linking point mutations in graSR and vraSR to the stepwise development of the VISA phenotype [46] increases the possibility of the future development of molecular methods for detection of hVISA strains

Using both Etest and broth microdilution techniques, our study found an association between higher vancomycin MIC values and presence of the hVISA phenotype. This finding is consistent with a recent report from Detroit [6]. Several studies have reported higher rates of treatment failure [47–49] and mortality [50] among patients with infections caused by MRSA isolates with higher vancomycin MIC values. In the current study, the mortality associated with infections caused by isolates with a vancomycin MIC of 2 mg/L by Etest was particularly high (66%), although the difference did not achieve statistical significance. This finding agrees with that of a recent study demonstrating suboptimal cure rates for vancomycin in an animal model of MRSA aortic endocarditis with a high vancomycin MIC (2 mg/L) [8]

The present study has limitations. Our sample size was relatively small, limiting our ability to detect significant differences between VSSA and hVISA infections or between infections with different vancomycin MICs. Next, we had no information regarding timing of complications, antibiotic dosing, trough serum concentration monitoring, antibiotic therapy before the onset of MRSA bacteremia, or the ultimate duration of bacteremia after therapy was initiated. We were also unable to consider why individual patients did or did not undergo cardiac surgery or why patients with VSSA infection underwent cardiac surgery twice as frequently as those with hVISA (30.4% vs 15.8%; P=.353). Standard practice in the ICE-PCS Microbiological Repository is to evaluate the first available bloodstream isolate. However, patients transferred from other facilities may have had bacteremia for extended periods of time before arrival at the ICE study center. Finally, because serial isolates from the same patient were not routinely available, we were unable to further evaluate the association between persistent bacteremia and the hVISA phenotype

In summary, this study makes several key observations. The hVISA phenotype occurred in more than one-quarter of MRSA IE isolates and was associated with a higher frequency of IE complications, such as persistent bacteremia and heart failure. hVISA was geographically distributed, with the highest rates being encountered in MRSA IE isolates from Europe or Oceania. hVISA and VSSA isolates were genotypically similar in this study. Vancomycin MIC values of 2 mg/L were relatively uncommon but were associated with hVISA. Future studies are required to validate the findings of this investigation, to clarify specific risk factors for the emergence of hVISA, and to identify practical, effective screening tests to detect hVISA in the clinical setting

ICE Microbiology Investigator Index 2009

Australia—Eugene Athan, Owen Harris (Barwon Health); Tony M. Korman (Monash Medical Centre); Despina Kotsanas (Southern Health); Phillip Jones, Porl Reinbott, Suzanne Ryan (University of New South Wales). Brazil—Claudio Querido Fortes (Hospital Universitário Clementino Fraga Filho/Universidade Federal do Rio de Janeiro). Chile—Patricia Garcia, Sandra Braun Jones (Hospital Clínico Pontificia Universidad Católica de Chile); Bruno Barsic, Suzana Bukovski (University Hospital for Infectious Diseases). France—Christine Selton-Suty, Neijla Aissa, Thanh Doco-Lecompte (Centre Hospitalier Universitaire Nancy-Brabois); François Delahaye, François Vandenesch (Hôpital Louis Pradel); Pierre Tattevin (Pontchaillou University); Bruno Hoen, Patrick Plesiat (University Medical Center of Besançon). Greece—Helen Giamarellou, (Attikon University General Hospital). Italy—Emanuele Durante-Mangoni, Marie-Françoise Tripodi, Riccardo Utili, Roberta Casillo, Susanna Cuccurullo (II Università di Napoli); Pierre Yves Donnio, Annibale Raglio, Fredy Suter (Ospedali Riuniti di Bergamo). Lebanon—Souha S. Kanj (American University of Beirut Medical Center). New Zealand—Arthur Morris (Diagnostic Medlab), David R. Murdoch (University of Otago). Slovenia—Manica Mueller Premru, Tatjana Lejko-Zupanc (Medical Center Ljubljana). Spain—Ana del Rio, Asuncion Moreno, Carlos Paré, Carlos Falces, Carlos Cervera, Carlos A. Mestres, Cristina Garcia-de-la-Maria, Elisa De Lazzari, Francesc Marco, Jose M. Gatell, Jose Ramirez, José M. Miró, Magda Heras, Manuel Azqueta, Manuel Almela, Maria-Jesús Jiménez-Expósito, Marta Sitges, Natividad de Benito, Salvador Ninot, Xavier Claramonte, Ximena Castañeda, Yolanda Armero (Hospital Clinic Institut d’Investigacions Biomèdiques August Pi i Sunyer, University of Barcelona); Emilio Bouza, Marta Rodríguez-Créixems, Victor Ramallo (Hospital General Universitario Gregorio Marañón). United States—Suzanne Bradley (Ann Arbor VA Medical Center); Dannah Wray, Lisa Steed, Robert Cantey (Medical University of South Carolina); Gail Peterson (University of Texas Southwestern Medical Center); Christopher W. Woods, G. Ralph Corey, L. Barth Reller, Vance G. Fowler, Jr, Vivian H. Chu (Duke University Medical Center)

ICE Coordinating Center. Eric Tsung-Cheng Hsieh, Judy Stafford, Khaula Baloch, G. Ralph Corey, Thomas Redick, Tina Harding, Vance G. Fowler, Jr, Vivian H. Chu

ICE Publications Committee. Andrew Wang, Arnold S. Bayer, A. W. Karchmer, Bruno Hoen, Christopher H. Cabell, Daniel J. Sexton, David R. Murdoch, Eugene Athan, José M. Miró, G. Ralph Corey, Kevin Anstrom, Vance G. Fowler, Jr, Vivian H. Chu

ICE Steering Committee. Arnold S. Bayer, Bruno Hoen, Christopher H. Cabell, Daniel J. Sexton, David T. Durack, Didier Raoult, Ethan Rubinstein, G. Ralph Corey, José M. Miró, Lars Olaison, Phillipe Moreillon, Susannah Eykyn, Vance G. Fowler, Jr, Vivian H. Chu

Potential Conflicts of Interest

Potential conflicts of interest: V.G.F. has served as a consultant for Astellas, Cubist, Inhibitex, Merck, Johnson & Johnson, and Leo Pharmaceuticals; reports having received grant or research support from Astellas, Cubist, Merck, Theravance, Inhibitex, Cerexa, and the National Institutes of Health; reports having received honoraria from Arpida, Astellas, Cubist, Inhibitex, Merck, Pfizer, Targanta, Theravance, and Ortho-McNeil; and has served on an advisory committee and on a speakers' bureau for Cubist. G.R.C. serves as a consultant for Theravance, Cubist, AstraZeneca Astellas, Cerexa, Merck, Cempra, Pfizer, Arpida, GlaxoSmithKline, Inimex, Targanta, and Trius (payment donated to the Global Health Fund at Duke University) and receives research contracts or grants from Theravance, Innocoll, Cerexa, and Cempra (funds provided through Duke University to support research or salary). S.S.K. is on the advisory board for Sanofi Aventis. M.J.R. is a speaker for Cubist, Wyeth, Pfizer, Astellas, Theravance Forest, Ortho-McNeil, and Targanta. C.W.W. has served as a consultant for Roche Molecular, reports having received grant or research support from Cubist and Roche Molecular, has served on advisory committees for bioMerieux and Astellas, and has been involved in clinical trials for bioMerieux and Cepheid. All other authors report no potential conflicts

Presented in part: 48th Interscience Conference on Antimicrobial Agents and Chemotherapy/46th Annual Meeting of the Infectious Diseases Society of America, Washington, DC, 25–28 October 2008 (abstract C1-197)

Financial support: Multiplex polymerase chain reaction and pulsed-field gel electrophoresis were funded by Cubist Pharmaceuticals (grant to C.W.W.). Population analyses were funded by Astellas (MRSA Center of Excellence grant to V.G.F., administered by Fallon Medica). Multilocus sequence typing were funded by the National Institutes of Health (grant R01-AI059111 to V.G.F.). The International Collaboration on Endocarditis Coordinating Center acknowledges the generous support of investigator donations from Bruno Barsic, G.R.C., V.G.F., David Gordon, and Andrew Wang and educational grants from Cubist Pharmaceuticals and the International Society of Cardiovascular Infectious Diseases