Adult methicillin-resistant Staphylococcus aureus bacteremia in Taiwan: clinical significance of non–multi-resistant antibiogram and Panton–Valentine leukocidin gene

Available online at www.sciencedirect.com

us Disease 59 (2007) 365–371www.elsevier.com/locate/diagmicrobio

Diagnostic Microbiology and Infectio

Adult methicillin-resistant Staphylococcus aureus bacteremia in Taiwan:clinical significance of non–multi-resistant antibiogram and

Panton–Valentine leukocidin geneJiun-Ling Wanga, Jann-Tay Wanga, Shey-Ying Chenb, Po-Ren Hsueha,c, Hsiang-Chi Kunga,d,

Yee-Chun Chena, Shan-Chwen Changa,e,⁎aDepartment of Internal Medicine, National Taiwan University Hospital, Taipei 100, Taiwan

bEmergency Medicine, National Taiwan University Hospital, Taipei 100, TaiwancLaboratory Medicine, National Taiwan University, Taipei 100, Taiwan

dDepartment of Internal Medicine, National Taiwan University Hospital Yun-Lin Branch, Douliou, TaiwaneGraduate Institute of Clinical Pharmacy, College of Medicine, National Taiwan University, Taipei 100, Taiwan

Received 11 January 2007; accepted 27 June 2007

Abstract

It is poorly defined whether or not adult patients with methicillin-resistant Staphylococcus aureus (MRSA) bacteremia with a non–multi-resistant antibiogram phenotype and Panton–Valentine leukocidin (PVL) gene carriage have different clinical syndromes. Clinicalcharacteristics of 95 adult patients of MRSA bacteremia, with isolates that were non–multi-resistant to non–β-lactam, were compared with acontemporaneous multiresistant group. Independent risk factors other than community-associated MRSA bacteremia patients associated withrecovery of non–multi-resistant MRSA isolates by multivariate analysis included deep-seated infection and catheter insertion site infection.Older age, intensive care unit-onset bacteremia, and postoperative infection were negative independent risk factors associated with non–multi-resistant MRSA isolates. Most of the 60 recoverable non–multi-resistant MRSA isolates belonged to multilocus sequence type 59, and allisolates belonged to staphylococcal chromosomal cassette mec (SCCmec) element type IV or type V. Most PVL-positive MRSA isolatesbelonged to SCCmec V. PVL-positive CA-MRSA isolates could cause more deep-seated infections in patients presented with non–multi-resistant MRSA bacteremia.© 2007 Elsevier Inc. All rights reserved.

Keywords: Methicillin-resistant Staphylococcus aureus; Toxin; Bacteremia; Community-acquired infection

1. Introduction

Since the late 1990s, community-associated methicillin-resistant Staphylococcus aureus (CA-MRSA) has emergedworldwide in patients who do not display health care-associated risk factors for MRSA infection (Naimi et al.,2003; Vandenesch et al., 2003). The accepted definition ofhealth care-associatedMRSA (HA-MRSA) infection includesMRSA infection with the presence of any of the followinghealth care-associated risk factors within 1 year before theindex MRSA bacteremia culture: 1) receipt of systemicantimicrobial treatment, 2) residence in a long-term care

⁎ Corresponding author. Tel.: +886-2-23123456x5401; fax: +886-2-23958721.

E-mail address: [email protected] (S.-C. Chang).

0732-8893/$ – see front matter © 2007 Elsevier Inc. All rights reserved.doi:10.1016/j.diagmicrobio.2007.06.021

facility, 3) prior hospitalization to an acute care facility, 4) useof central intravenous catheters or long-term venous accessdevices, 5) use of urinary catheters, 6) use of other long-termpercutaneous devices, 7) prior surgical procedures, and/or,8) need of dialysis (Naimi et al., 2003; Seybold et al., 2006).Traditional nosocomial MRSA strains are multiresistant tonon–β-lactam antibiotic, and the emerging CA-MRSA strainsare usually susceptible to various antibiotics (Naimi et al.,2003; Vandenesch et al., 2003). However, the CA-MRSAstrains in Taiwan are only susceptible to trimethoprim/sulfamethoxazole, ciprofloxacin, gentamicin, and mino-cycline (Boyle-Vavra et al., 2005), and have an extraordinarilyhigh resistance rate to erythromycin and clindamycin (Fanget al., 2004;Wang et al., 2004; Boyle-Vavra et al., 2005; Chenet al., 2005). It has been reported that CA-MRSA strain mayspread in the hospital and cause nosocomial infections (Healy

366 J.-L. Wang et al. / Diagnostic Microbiology and Infectious Disease 59 (2007) 365–371

et al., 2004; Huang et al., 2006; Seybold et al., 2006). There-fore, it would become difficult to differentiate nosocomialMRSA infection from CA-MRSA infection only by theantibiogram of the isolates once the CA-MRSA strains areintroduced into and spreading in a hospital setting.

In addition to the difference of antibiogram betweenCA-MRSA strains and nosocomialMRSA strains, CA-MRSAstrains have a different type of staphylococcal chromosomalcassette mec (SCCmec) element gene compared with nosoco-mial MRSA isolates (Naimi et al., 2003; Vandenesch et al.,2003). SCCmec V is found in an area in Taiwan, which is anovel SCCmec variant found in CA-MRSA strains and isdifferent from SCCmec IV, which is the most common typeof CA-MRSA isolates found in other countries (Vandeneschet al., 2003; Boyle-Vavra et al., 2005). The SCCmec V CA-MRSA strains are usually Panton–Valentine leukocidin (PVL)positive according to previous reports from Taiwan (Chenet al., 2005). However, the clinical manifestations of PVL-producing CA-MRSA–harboring SCCmec V in adult MRSAbacteremia patients are not well described.

Because most CA-MRSA bloodstream infections and theirclinical and microbiologic characteristic are reported inchildren (Gonzalez et al., 2005; Bocchini et al., 2006), it isnot known if the typical clinical characteristics of CA-MRSAinfection in children could be applied to adult patients.Whetheradult patients with non–multi-resistant MRSA bacteremiaimply CA-MRSA bacteremia and would have similar clinicalpresentation as those in pediatric patients is not known.

The purposes of this study are to illustrate the clinical andepidemiologic features of adult patients with non–multi-resistant bacteremia in comparison with those of multi-resistant MRSA bacteremia and to evaluate the role of PVLgenes in CA-MRSA and HA-MRSA isolates.

2. Materials and methods

2.1. Case selection and definitions

This study was approved according to the local institutionalethics review board requirements and was performed atNational Taiwan University Hospital (NTUH), Taipei, Taiwan,a 2400-bed teaching hospital in northern Taiwan, whichprovides both primary and tertiary care. The nosocomialbloodstream infection rate of S. aureus at NTUH in 2005 wasabout 23.9 people per 10000 discharges, and the prevalence ofMRSA in nosocomial S. aureus isolates from bloodwas 72.9%.

Because there has been no reported differences betweenCA-MRSA and HA-MRSA in the erythromycin andclindamycin resistance rates in Taiwan (Fang et al., 2004;Wang et al., 2004), we identified from our microbiologydatabase the adult MRSA bacteremia patients with isolatessusceptible to trimethoprim/sulfamethoxazole, gentamicin,and minocycline. We defined isolates with this phenotype asnon–multi-resistant MRSA, which is the usual antibiogramphenotype of CA-MRSA isolates in Taiwan (Fang et al.,2004; Wang et al., 2004).

In vitro antimicrobial susceptibility testing of MRSAisolates in our institution was done by disk diffusion methodin accordance with Clinical and Laboratory Standards Institute(formerly National Committee for Clinical Laboratory Stan-dards [NCCLS]) break point criteria (NCCLS, 2005). FromJanuary 1, 2002, through December 31, 2005, 95 nonduplicateand consecutive adult patients older than 18 years with non–multi-resistant MRSA bacteremia were identified as the casegroup. One-to-one matched control MRSA bacteremiapatients with multiresistant isolates (resistant to 1 or more ofthe following 3 non–β-lactam antibiotics: trimethoprim/sulfamethoxazole, gentamicin, or minocycline) matched toeach case by calendar time within the same week wereretrospectively selected as the control group. Demographicdata, clinical syndromes, and outcomes were collected frompatient charts and used to compare the 2 groups.

HA-MRSA infection represents the presence of any ofhealth care-associated risk factors within 1 year before theindex MRSA bacteremia culture (Naimi et al., 2003; Seyboldet al., 2006). Nosocomial MRSA bacteremia was defined byan index MRSA-positive culture of blood obtained N48 hafter admission. Community-onset MRSA bacteremia wasdefined by anMRSA-positive culture of blood obtained froman outpatient or from an inpatient within 48 h of admission.CA-MRSA infection was defined by a community-onsetinfection in the absence of any of the health care-associatedrisk factor described above.

Catheter-related bacteremia was defined by a semiquanti-tative culture of the vascular catheter tip yielding greater than15 MRSA colonies in the absence of other sources ofbacteremia (Mermel et al., 2001). Catheter insertion siteinfection was defined as any sign of local infection(induration, erythema, heat, pain, and purulent drainage) atthe catheter insertion site. The diagnosis of infectiveendocarditis required those symptoms and signs to meet themodified Duke criteria (Li et al., 2000). Necrotizingpneumonia would display a cavity upon imaging of thechest. The other sites of infection at the onset of bacteremiawere defined according to the US Centers for Disease Controland prevention criteria (Garner et al., 1988). The diagnosis ofdeep-seated infection that required prolonged (4–6 weeks ormore) antibiotic therapy included complicated soft tissueinfection (pyomyositis, epidural abscess, and necrotizingfasciitis), osteomyelitis, septic arthritis, intravascular infec-tion, empyema, or abscess in any other organ such as eye,brain, lung, liver, or kidney. These were obtained according tothe clinical symptoms/signs—laboratory findings, positiveradiologic imaging, and positive culture from a normallysterile site—similar to the definition of complicated S. aureusinfection (Fowler et al., 2003).

2.2. Molecular typing and antimicrobial susceptibilitytesting

To illustrate whether isolates of non–multi-resistantantibiogram phenotype belonging to the CA-MRSA strain,defined as isolates carrying SCCmec IV or V, we genotyped

Table 1Univariate and multivariate analyses of patient characteristics in the non–multi-resistant and multiresistant groups

Patient characteristic Non–multi-resistantMRSA (n = 95)

MultiresistantMRSA (n = 95)

Crude OR(95% CI)

Adjusted OR(95% CI)

Age in years (mean ± SD) 60.6 ± 18.8 66.6 ± 16.4 0.98 (0.97–1.00) 0.98 (0.96–1.00)No. of males (% of total) 59 (62.1) 67 (70.5) 1.46 (0.80–2.67)Community onset 49 (51.6) 19 (20.0) 4.26 (2.24–8.11)Onset in the ICU 10 (10.5) 37 (38.5) 0.18 (0.09–0.40) 0.33 (0.14–0.82)CA-MRSA 25 (26.3) 0 (0)No systemic disease history 13 (13.7) 1 (1.1) 14.90 (1.91–116.39)Infection fociSkin and soft tissue 18 (18.9) 15 (15.8) 1.25 (0.59–2.65)Urinary tract 3 (3.2) 4 (4.2) 0.74 (0.16–3.41)Central catheter-related 23 (24.2) 35 (36.8) 0.55 (0.29–1.03)Catheter insertion site 10 (10.5) 3 (3.2) 3.61 (0.96–13.55) 7.24 (1.61–32.65)Postoperative wound 5 (5.3) 13 (13.7) 0.35 (0.12–1.03) 0.14 (0.03–0.57)Deep-seated infection 28 (29.5) 8 (8.4) 4.55 (1.95–10.61) 3.47 (1.10–10.93)Necrotizing pneumonia/empyema 9 (9.5) 1 (1.1) 9.84 (1.22–79.26)Endocarditis/mycotic aneurysm 9 (9.5) 3 (3.2) 3.21 (0.84–12.25)Bone and joint 10 (10.5) 4 (4.2) 2.68 (0.81–8.86)

PVL-positive isolates 32/60 (53.3) 0/40 (0)

Unless otherwise noted, data are number (%) of cases. CA-MRSA, no health care-associated risk factor within 1 year. Crude OR, OR in univariate analysis.Adjusted OR, OR in multivariable analysis of patient characteristics for predicting non–multi-resistant MRSA infection by comparing non–multi-resistant andmultiresistant MRSA bacteremia patients with a logistic regression model excluding CA-MRSA infection patients.

able 2A-MRSA versus community-onset HA-MRSA versus nosocomial MRSAacteremia patients in 95 cases of a non–multi-resistant MRSA bacteremia

linical syndromend outcome

CA-MRSAinfection,n = 25

Community-onsetHA-MRSAinfection,n = 24

NosocomialMRSAinfection,n = 46

P

VL-positiveisolates

11/13 (84.6) 9/19 (47.4) 12/28 (42.9) 0.037

fection fociSuperficial skinand soft tissue

4 (16.0) 3 (12.0) 3 (6.5) 0.432

Pneumonia(non-necrotizing)

1 (4.0) 3 (12.5) 4 (8.7) 0.064

Urinary tract 3 (12.0) 0 (0) 0 (0) 0.017Centralcatheter-related

0 6 (25.0) 17 (37.0) 0.002

Catheter insertionsite infection

0 2 (8.3) 8 (17.4) 0.058

Postoperativewound

0 2 (8.3) 3 (6.5) 0.321

eep-seatedinfection

15 (60.0) 8 (33.3) 5 (10.9) b0.001

Endocarditis/mycoticaneurysm

4 (16.0) 3 (12.5) 2 (4.3) 0.234

367J.-L. Wang et al. / Diagnostic Microbiology and Infectious Disease 59 (2007) 365–371

available nonduplicate blood isolates from the clinicalmicrobiology laboratory by multilocus sequence typing(MLST) and pulsed-field gel electrophoresis (PFGE)demonstrations of protein A gene (spa) and SCCmec.Susceptibilities to erythromycin, clindamycin, gentamicin,amikacin, ciprofloxacin, levofloxacin, tetracycline, minocy-cline, trimethoprim/sulfamethoxazole, rifampin, linezolid,and vancomycin were tested using the disc diffusion method(NCCLS, 2005). The presence of the PVL gene lukF-lukSwas investigated by polymerase chain reaction using apreviously described primer (Gillet et al., 2002; Lina et al.,1999). Presence of the SCCmec elements (I–V) and themecgene were determined as previously described (Ito et al.,2001; Ito et al., 2004; Ma et al., 2002). MLSTwas performedas previously described (Enright et al., 2000). Thepolymorphic X-region of the protein A gene (spa) wasanalyzed as previously described (Harmsen et al., 2003).

2.3. Statistical analyses

Statistical analyses of category variables were performedby the χ2 and Fisher's exact tests. Continuous variables werecompared by the Student t test. Statistical analyses wereperformed with SAS version 8.0 software (SAS, Cary, NC).

Bone and joint 5 (20) 3 (12.5) 2 (4.3) 0.114Necrotizingpneumonia/empyema

7 (28.0) 2 (8.3) 0 (0) 0.001

utcomePersistedbacteremiaN 7 days

4 (16.0) 2 (8.3) 4 (8.7) 0.447

Death within48 h

4 (16.0) 5 (20.8) 6 (13.0) 0.697

Death within14 days

4 (16.0) 5 (20.0) 15 (32.6) 0.259

3. Results

3.1. Patient and infection characteristics

From January 1, 2002, through December 31, 2005, a totalof 795 nonduplicate consecutive adult MRSA bacteremiapatients were identified at NTUH. The non–multi-resistantantibiogram phenotype constituted 95 (11.9%) of all adultMRSA bacteremia cases. The proportion of the non–multi-

TCb

Ca

P

In

D

O

368 J.-L. Wang et al. / Diagnostic Microbiology and Infectious Disease 59 (2007) 365–371

resistant antibiogram phenotype in the MRSA bacteremiacases increased gradually during the study period (2002,7.7%; 2003, 11.6%; 2004, 13.5%; 2005, 15.5%; P = 0.092 byχ2 test). Among these 95 cases, 25 were CA-MRSAinfections, 24 were community-onset HA-MRSA infections,and 46 were nosocomial MRSA infections. In contrast, all95 cases in the multiresistant control group were HA-MRSAinfections (19 cases of community-onset HA-MRSA infec-tion and 76 nosocomial infection cases), and none matchedthe definition of a CA-MRSA infection.

3.2. Non–multi-resistant versus multiresistantMRSA bacteremia

The demographics and clinical features of MRSAbacteremia patients with the non–multi-resistant and multi-resistant antibiograms are shown in Table 1. MRSAbacteremia patients with non–multi-resistant MRSA antibio-gram were less likely to be older than patients withmultiresistant MRSA (crude odds ratio [OR] per year, 0.98;95% confidence interval [CI], 0.97–1.00). In patients withnon–multi-resistant MRSA isolates, there were more cases ofcommunity-onset infection (crude OR, 4.26; 95% CI, 2.24–8.11) and cases having no history of systemic disease (crudeOR, 14.90; 95% CI, 1.91–116.39), but fewer cases with the

Fig. 1. PFGE patterns and phylogenetic tree of MRSA isolates with SCCmec type It437. All isolates with SCCmec V in this figure carried PVL genes.

index culture obtained in the intensive care unit (ICU) (crudeOR, 0.18; 95% CI, 0.09–0.40), compared with the multi-resistant group (Table 1). Concerning the clinical syndromes,there were more cases of deep-seated infection (crude OR,4.55; 95%CI, 1.95–10.61), necrotizing pneumonia/empyema(crude OR, 9.84; 95% CI, 1.22–79.26), and catheter insertionsite infection (crude OR, 3.61; 95% CI, 0.96–13.55), butfewer cases of postoperative wound infection (crude OR,0.35; 95% CI, 0.12–1.03), in the non–multi-resistant groupthan in the multiresistant group (Table 1).

Because all 25 cases of CA-MRSA infection involvednon–multi-resistant isolates and because CA-MRSA infec-tion is a well-known factor associated the non–multi-resistant isolates, a logistic regression model that excludedCA-MRSA infection was generated (Table 1). This modelfound deep-seated infection (adjusted OR, 3.47; 95% CI,1.10–10.93) and catheter insertion site infection (adjustedOR, 7.24; 95% CI, 1.61–32.65) as independent risk factorsfor recovery of non–multi-resistant MRSA from bloodculture. Older age (adjusted OR per year, 0.98; 95% CI,0.96–1.00), index culture in the ICU (adjusted OR, 0.33;95% CI, 0.14–0.82), and postoperative wound infection(adjusted OR, 0.14; 95% CI, 0.03–0.57) were significantlyless likely to have non–multi-resistant MRSA recovered.

Vor V. All the isolates shown were multilocus sequence type 59 and spa type

369J.-L. Wang et al. / Diagnostic Microbiology and Infectious Disease 59 (2007) 365–371

3.3. CA-MRSA versus community-onset HA-MRSA versusnosocomial HA-MRSA bacteremia

The comparisons of the clinical syndromes and outcomesbetween 25 CA-MRSA cases, 24 community-onsetHA-MRSA cases, and 46 nosocomial HA-MRSA cases of95 non–multi-resistant MRSA bacteremia patients are shownin Table 2. More instances of necrotizing pneumonia/empyema, urinary tract infection, and deep-seated infection,but fewer instances of catheter-related infections, were seenin CA-MRSA bacteremia patients than in community-onsetHA-MRSA and nosocomial MRSA bacteremia patients(P b 0.05).

3.4. Genotyping of MRSA isolates and antimicrobialsusceptibility test

Sixty nonduplicated isolates of non–multi-resistantMRSA from the 95 cases were recovered for furthermicrobiologic study. The demographic features and clinicalsymptoms did not differ between patients with or withoutisolates available for genotyping (data not shown). These60 non–multi-resistant MRSA isolates displayed goodsusceptibility rates to minocycline (100%), trimethoprim/sulfamethoxazole (100%), gentamicin (100%), rifampin(100%), linezolid (100%), vancomycin (100%), levofloxacin(98.3), ciprofloxacin (93.3), and amikacin (85%), but poorsusceptibility rates to erythromycin (10%), clindamycin(8.3%), and tetracycline (31.7%).

Table 3SCCmec type and demographic data in PVL-positive and PVL-negativeMRSA bacteremia patients

Demographic andmicrobiologicfeature

PVL-positiveCA-MRSA(n = 11)

PVL-positiveHA-MRSA(n = 21)

PVL-negativeHA-MRSA(n = 26)

P

SCCmec V 9 (81.8) 20 (95.2) 2 (7.7) b0.001Sex, male 8 (72.7) 12 (57.1) 16 (61.5) 0.687Elderly (ageN65 years old)

3 (27.3) 14 (66.7) 12 (46.2) 0.093

No underlyingdisease

7 (63.6) 1 (4.8) 0 (0) b0.001

Infection siteSuperficial skinand soft tissue

2 (18.2) 2 (9.5) 2 (7.7) 0.625

Deep-seatedinfection

9 (81.8) 5 (23.8) 6 (23.1) 0.001

Necrotizingpneumonia/empyema

5 (45.5) 1 (4.8) 1 (3.8) b0.01

OutcomeDeath within48 h

1 (9.1) 4 (19.0) 4 (15.4) 0.761

Death within14 days

1 (9.1) 5 (23.8) 6 (23.1) 0.572

Unless otherwise noted, data are number (%) of cases. Data from 2 patientswith PVL-negative CA-MRSA infection are not shown. HA-MRSAinfection included community-onset HA-MRSA and nosocomialMRSA infection.

Most non–multi-resistant MRSA isolates belonged toST59 (n = 56, 93.3%) and spa type t437 (n = 55, 91.7%). All60 non–multi-resistant isolates carried mecA gene andbelonged to SCCmec IV (n = 29, 48.3%) or SCCmec V(n = 31, 51.7%). There were 53.3% (32/60) of non–multi-resistant MRSA that carried the PVL gene. We alsoperformed genotyping on 40 multiresistant MRSA isolates;none carried the PVL gene or SCCmecV. SCCmecV isolatesharbored the PVL gene significantly more often than didSCCmec IV isolates (93.5% versus 10.3%, P b 0.05), andSCCmec IV isolates had a different pulsotype in comparisonwith those of SCCmec V (Fig. 1). Approximately 60% ofMRSA isolates with the SCCmec V gene belonged to themajor pulsotype (coefficient of similarity N 70%) (Fig. 1).

Compared with PVL-positive HA-MRSA patientsand PVL-negative HA-MRSA patients, PVL-positiveCA-MRSA patients were more likely to have no underlyingsystemic disease, and more likely to have deep-seatedinfection, and necrotizing pneumonia/empyema (P b 0.05)(Table 3). SCCmec V was more common in PVL-positiveHA-MRSA or PVL-positive CA-MRSA groups than in thePVL-negative HA-MRSA group (P b 0.05) (Table 3). Therewas no statistically significant difference in the mortality inthese 3 groups.

4. Discussion

In this case-control study, we report the clinical andepidemiologic characteristics of MRSA bacteremia patientswith non–multi-resistant antibiogram. We report thatCA-MRSA strains can also cause bloodstream infections ina health care-associated setting, similar to a recent reportfrom a US urban public hospital (Seybold et al., 2006).Although our cases were in older people with morecomorbidities and the CA-MRSA isolates were not thesame in MLST typing as epidemic strain in other parts of theworld, our patients presented with similar clinical features ofsevere deep-seated infections as found in pediatric patients(Bocchini et al., 2006; Gonzalez et al., 2005).

Our study corroborates previous findings concerningCA-MRSA infections in Taiwan, which documented ST59,with SCCmec Vas the predominant CA-MRSA strain (Chenet al., 2005; Wang et al., 2004). The ST59 strain ofCA-MRSA has been reported in Australia (Gosbell et al.,2006), Europe (Fossum and Bukholm, 2006), and SanFrancisco (Pan et al., 2005) but was not a common strain intheir CA-MRSA isolates. CA-MRSA strains in Taiwan differfrom those found in the United States and Europe(Vandenesch et al., 2003) in that SCCmecV, but not SCCmecIV, is more likely to carry the virulent factor PVL. Ourfindings extend the knowledge of the clinical profile ofCA-MRSA strains carrying the PVL gene. SCCmec V andPVL-positive strains related to necrotizing pneumonia havealso been reported recently in Greece (Gerogianni et al.,2006). Most of our PVL-positive cases were community-

370 J.-L. Wang et al. / Diagnostic Microbiology and Infectious Disease 59 (2007) 365–371

associated or community-onset HA-MRSA infections.Although PVL was initially thought to be a good markerfor CA-MRSA infection (Vandenesch et al., 2003), it wasrecently reported that 30% of isolates from health care-associated MRSA bacteremia cases carried the PVL gene(Seybold et al., 2006). Our study showed 45% of isolatesfrom HA-MRSA bacteremia cases that carried the PVL geneThe impact of CA-MRSA strains harboring this gene on bothcommunity- and health care-associated infections needs to beclarified (Etienne, 2005; Zetola et al., 2005). Other thanfolliculitis and necrotizing pneumonia, (Francis et al., 2005),some rare disease entities such as pyomyositis, bone and jointinfections, and intravascular infection have been reported inPVL-positive MRSA infections in children and adolescents(Bocchini et al., 2006; Gonzalez et al., 2005; Pannaraj et al.,2006). Our case series also indicates that severe life-threatening infections such as mycotic aneurysm, endocardi-tis, pyomyositis, and other deep-seated infections may alsooccur in adult CA-MRSA infection patients.

The present study demonstrates the increased presence ofdeep-seated community onset infections without health care-associated risk factors. PVL-positive CA-MRSA also hadmore deep-seated infection than PVL-positive HA-MRSAbacteremia patients. One previous staphylococcal bacteremiastudy that included fewer cases of MRSA infection foundthat community acquisition is one predictor of complicatedinfections (Fowler et al., 2003). A preliminary comparativestudy of CA-MRSA and HA-MRSA bacteremia patients inthe United States showed that CA-MRSA bacteremiapatients are more likely to have deep-seated infections thanHA-MRSA bacteremia patients (Ray et al., 2005). Thisdifferent clinical syndrome may be explained by differentpresentation times and underlying diseases betweenCA-MRSA cases and HA-MRSA cases.

Although different definitions of non–multi-resistantMRSA were used, our non–multi-resistant MRSA isolateshave a predisposition to affect the young and relativelyhealthy people in the community setting compared withmultiresistant isolates. This finding is similar to those foundin previous studies (Gosbell et al., 2001; Naimi et al., 2003).More deep-seated infections, but less index culture in theICU and less postoperative wound infection, were found inthe non–multi-resistant group.

From our study, it was found that the susceptibility oferythromycin, clindamycin, and tetracycline could not beused to differentiate CA-MRSA strains from HA-MRSAstrains. The reason why we had such a high rate of resistanceto erythromycin and clindamycin in Taiwan has beendiscussed in previous studies (Boyle-Vavra et al., 2005;Wang et al., 2004). The prevalence of tetracycline resistancecontinues to increase in CA-MRSA strains, and there werealso reports of increasing circulation of CA-MRSA strainswith resistance to clindamycin, tetracycline, and othernon–β-lactams in Europe and the United States (Pan et al.,2005; Ramdani-Bouguessa et al., 2006). Therefore, weshould be careful in using any non–β-lactam-susceptible

antibiogram as the screening tool of CA-MRSA strains,especially if the resistant rate changed rapidly.

The limitation of this study is that it is retrospective andnot all isolates were available for genotyping and virulencestudies. Possible selection bias in the genotyping studyexisted. In addition, the unique definition of non–multi-resistant MRSA using local CA-MRSA antibiotic phenotypein Taiwan may not be applicable to other countries.

In conclusion, younger age, community acquisition,and deep-seated infection were associated with non–multi-resistant antibiogram in adult patients with MRSA bacter-emia. SCCmec V and PVL-positive strain was a commonstrain in CA-MRSA bacteremia in Taiwan. PVL-positiveCA-MRSA isolates could cause more deep-seatedinfections in patients presented with non–multi-resistantMRSA bacteremia.

Acknowledgments

The authors thank Fu-Chang Hu for his contribution onstatistics.

This study was supported by grants from the NationalScience Council in Taiwan (NSC 94-2314-B-002-184).

References

Bocchini CE, Hulten KG, Mason Jr EO, Gonzalez BE, HammermanWA, Kaplan SL (2006) Panton–Valentine leukocidin genes areassociated with enhanced inflammatory response and local disease inacute hematogenous Staphylococcus aureus osteomyelitis in children.Pediatrics 117:433–440.

Boyle-Vavra S, Ereshefsky B, Wang CC, Daum RS (2005) Successfulmultiresistant community-associated methicillin-resistant Staphylococ-cus aureus lineage from Taipei, Taiwan, that carries either the novelstaphylococcal chromosome cassettemec (SCCmec) type VTor SCCmectype IV. J Clin Microbiol 43:4719–4730.

Chen FJ, Lauderdale TL, Huang IW, Lo HJ, Lai JF, Wang HY, Shiau YR,Chen PC, Ito T, Hiramitsu K (2005) Methicillin-resistant Staphylococ-cus aureus in Taiwan. Emerg Infect Dis 11:1760–1763.

Enright MC, Day NP, Davies CE, Peacock SJ, Spratt BG (2000) Multilocussequence typing for characterization of methicillin-resistant andmethicillin-susceptible clones of Staphylococcus aureus. J Clin Micro-biol 38:1008–1015.

Etienne J (2005) Panton–Valentine leukocidin: a marker of severity forStaphylococcus aureus infection? Clin Infect Dis 41:591–593.

Fang YH, Hsueh PR, Hu JJ, Lee PI, Chen JM, Lee CY, Huang LM (2004)Community-acquired methicillin-resistant Staphylococcus aureus inchildren in northern Taiwan. J Microbiol Immunol Infect 37:29–34.

Fossum AE, Bukholm G (2006) Increased incidence of methicillin-resistantStaphylococcus aureus ST80, novel ST125 and SCCmecIV in the south-eastern part of Norway during a 12-year period. Clin Microbiol Infect12:627–633.

Fowler Jr VG, Olsen MK, Corey GR, Woods CW, Cabell CH, Reller LB,Cheng AC, Dudley T, Oddone EZ (2003) Clinical identifiers ofcomplicated Staphylococcus aureus bacteremia. Arch Intern Med163:2066–2072.

Francis JS, Doherty MC, Lopatin U, Johnston CP, Sinha G, Ross T, Cai M,Hansel NN, Perl T, Ticehurst JR, Carroll K, Thomas DL, NuermbergerE, Bartlett JG (2005) Severe community-onset pneumonia in healthyadults caused by methicillin-resistant Staphylococcus aureus carryingthe Panton–Valentine leukocidin genes. Clin Infect Dis 40:100–107.

371J.-L. Wang et al. / Diagnostic Microbiology and Infectious Disease 59 (2007) 365–371

Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM (1988)CDC definitions for nosocomial infections. Am J Infect Control16:128–140.

Gerogianni I, Mpatavanis G, Gourgoulianis K, Maniatis A, Spiliopoulou I,Petinaki E (2006) Combination of staphylococcal chromosome cassetteSCCmec type Vand Panton–Valentine leukocidin genes in a methicillin-resistant Staphylococcus aureus that caused necrotizing pneumonia inGreece. Diagn Microbiol Infect Dis 56:213–216.

Gillet Y, Issartel B, Vanhems P, Fournet JC, Lina G, Bes M, Vandenesch F,Piemont Y, Brousse N, Floret D, Etienne J (2002) Association betweenStaphylococcus aureus strains carrying gene for Panton–Valentineleukocidin and highly lethal necrotising pneumonia in young immuno-competent patients. Lancet 359:753–759.

Gonzalez BE, Martinez-Aguilar G, Hulten KG, Hammerman WA, Coss-BuJ, Avalos-Mishaan A, Mason Jr EO, Kaplan SL (2005) Severestaphylococcal sepsis in adolescents in the era of community-acquired methicillin-resistant Staphylococcus aureus. Pediatrics115:642–648.

Gosbell IB, Barbagiannakos T, Neville SA, Mercer JL, Vickery AM,O'Brien FG, Coombs GW, Malkowski MJ, Pearson JC (2006) Non-multiresistant methicillin-resistant Staphylococcus aureus bacteraemiain Sydney, Australia: emergence of EMRSA-15, Oceania, Queenslandand Western Australian MRSA strains. Pathology 38:239–244.

Gosbell IB, Mercer JL, Neville SA, Crone SA, Chant KG, Jalaludin BB,Munro R (2001) Non-multiresistant and multiresistant methicillin-resistant Staphylococcus aureus in community-acquired infections.MedJ Aust 174:627–630.

Harmsen D, Claus H, Witte W, Rothganger J, Claus H, Turnwald D, Vogel U(2003) Typing of methicillin-resistant Staphylococcus aureus in auniversity hospital setting by using novel software for spa repeatdetermination and database management. J Clin Microbiol41:5442–5448.

Healy CM, Hulten KG, Palazzi DL, Campbell JR, Baker CJ (2004)Emergence of new strains of methicillin-resistant Staphylococcus aureusin a neonatal intensive care unit. Clin Infect Dis 39:1460–1466.

Huang YC, Su LH, Wu TL, Lin TY (2006) Changing molecularepidemiology of methicillin-resistant Staphylococcus aureus blood-stream isolates from a teaching hospital in Northern Taiwan. J ClinMicrobiol 44:2268–2270.

Ito T, Katayama Y, Asada K, Mori N, Tsutsumimoto K, Tiensasitorn C,Hiramatsu K (2001) Structural comparison of three types of staphylo-coccal cassette chromosome mec integrated in the chromosome inmethicillin-resistant Staphylococcus aureus. Antimicrob Agents Che-mother 45:1323–1336.

Ito T, Ma XX, Takeuchi F, Okuma K, Yuzawa H, Hiramatsu K (2004) Noveltype V staphylococcal cassette chromosome mec driven by a novelcassette chromosome recombinase, ccrC. Antimicrob Agents Chemother48:2637–2651.

Li JS, Sexton DJ, Mick N, Nettles R, Fowler Jr VG, Ryan T, Bashore T,Corey GR (2000) Proposed modifications to the Duke criteria for thediagnosis of infective endocarditis. Clin Infect Dis 30:633–638.

Lina G, Piemont Y, Godail-Gamot F, Bes M, Peter MO, Gauduchon V,Vandenesch F, Etienne J (1999) Involvement of Panton–Valentineleukocidin-producing Staphylococcus aureus in primary skin infectionsand pneumonia. Clin Infect Dis 29:1128–1132.

Ma XX, Ito T, Tiensasitorn C, Jamklang M, Chongtrakool P, Boyle-Vavra S,Daum RS, Hiramatsu K (2002) Novel type of staphylococcal cassettechromosome mec identified in community-acquired methicillin-resistantStaphylococcus aureus strains. Antimicrob Agents Chemother46:1147–1152.

Mermel LA, Farr BM, Sherertz RJ, Raad II, O'Grady N, Harris JS, CravenDE, Infectious Diseases Society of America, American College ofCritical Care Medicine, Society for Healthcare Epidemiology ofAmerica (2001) Guidelines for the management of intravascularcatheter-related infections. Clin Infect Dis 32:1249–1272.

Naimi TS, LeDell KH, Como-Sabetti K, Borchardt SM, Boxrud DJ, EtienneJ, Johnson SK, Vandenesch F, Fridkin S, O'Boyle C, Danila RN,Lynfield R (2003) Comparison of community- and health care-associated methicillin-resistant Staphylococcus aureus infection. JAMA290:2976–2984.

National Committee for Clinical Laboratory Standards (2005) 2005 NationalCommittee for Clinical Laboratory Standards, performance standards forantimicrobial susceptibility testing. 14th Informational SupplementM100-S15 NCCLS. Wayne (PA).

Pan ES, Diep BA, Charlebois ED, Auerswald C, Carleton HA, SensabaughGF, Perdreau-Remington F (2005) Population dynamics of nasal strainsof methicillin-resistant Staphylococcus aureus—and their relation tocommunity-associated disease activity. J Infect Dis 192:811–818.

Pannaraj PS, Hulten KG, Gonzalez BE, Mason Jr EO, Kaplan SL (2006)Infective pyomyositis and myositis in children in the era of community-acquired, methicillin-resistant Staphylococcus aureus infection. ClinInfect Dis 43:953–960.

Ramdani-Bouguessa N, Bes M, Meugnier H, Forey F, Reverdy ME, Lina G,Vandenesch F, Tazir M, Etienne J (2006) Detection of methicillin-resistant Staphylococcus aureus strains resistant to multiple antibioticsand carrying the Panton–Valentine leukocidin genes in an Algiershospital. Antimicrob Agents Chemother 50:1083–1085.

Ray SM, Farley MM, Ladson JL, Craig AS, Lynfield R, Townes JM, Petit S,Harrison LH, Reingold A, Gershman K, Dumyati G, Morrison M,Klevens M (2005) Invasive methicillin-resistant Staphylococcus aureus(MRSA) infections. Comparison of community-acquired and healthcare-associated (HA) disease. Program and Abstracts of the InfectiousDiseases Society of America 43rd Annual Meeting; San Francisco,California. Abstract 682.

Seybold U, Kourbatova EV, Johnson JG, Halvosa SJ, Wang YF, King MD,Ray SM, Blumberg HM (2006) Emergence of community-associatedmethicillin-resistant Staphylococcus aureus USA300 genotype as amajor cause of health care-associated blood stream infections. ClinInfect Dis 42:647–656.

Vandenesch F, Naimi T, Enright MC, Lina G, Nimmo GR, Heffernan H,Liassine N, Bes M, Greenland T, Reverdy ME, Etienne J (2003)Community-acquired methicillin-resistant Staphylococcus aureus carry-ing Panton–Valentine leukocidin genes: worldwide emergence. EmergInfect Dis 9:978–984.

Wang CC, Lo WT, Chu ML, Siu LK (2004) Epidemiological typing ofcommunity-acquired methicillin-resistant Staphylococcus aureus iso-lates from children in Taiwan. Clin Infect Dis 39:481–487.

Zetola N, Francis JS, Nuermberger EL, Bishai WR (2005) Community-acquired methicillin-resistant Staphylococcus aureus: an emergingthreat. Lancet Infect Dis 5:275–286.