We selected 2 collections of strains with 151 nonduplicated MRSA isolates identified in patients or carriers treated at our institution. The first collection was from a retrospective review of laboratory records and included non–multidrug-resistant (gentamicin- and ciprofloxacin-susceptible) strains collected during 1993–2002 that had a phenotype different from the endemic healthcare-associated MRSA (HA-MRSA) strain in Geneva. The prevalent HA-MRSA clone in Geneva is sequence type (ST) 228-MRSA-I (CC5), which shows resistance to gentamicin, ciprofloxacin, clindamycin, and erythromycin. HA-MRSA strain ST8-MRSA-IV has been sporadically introduced from France. This strain has the same phenotype as ST228-MRSA-I (CC5) except for its susceptibility to gentamicin (5). We included all strains resistant to or with intermediate susceptibility to fusidic acid, a characteristic of many CA-MRSA isolates in Europe.

The second collection was isolates selected from patients prospectively identified as colonized or infected with CA-MRSA by the CA-MRSA surveillance program during 2003–2005 (6). CA-MRSA was defined as any isolate with an antimicrobial drug resistance profile different from the strain endemic in the Geneva healthcare setting and diagnosed in a patient without a history of hospitalization in the previous 12 months.

MRSA identification was performed by using standard methods (7) according to Clinical and Laboratory Standard Institute recommendations (8) and confirmed by quantitative PCR (9). Genomic DNA isolated from 1 colony was tested by quantitative PCR for staphylococcal cassette chromosome mec (SCCmec) elements, accessory gene regulator group, and the PVL gene (10,11). Presence of type V cassette, TSST-1, and exfoliatin toxins was assessed by using specific oligonucleotides (sequences are available at www.genomic.ch/sup6.php). Multiple-locus variable-number tandem repeat analysis, which consisted of a multiplex PCR with 10 primer pairs, and multilocus sequence typing were performed as reported (11,12).

Since late 2002, all patient demographic and epidemiologic data have been reviewed and recorded on a standardized form by a public health nurse (6). For this analysis, we included only those patients who were seen at our institution or outpatient clinic.

A total of 92 strains from 51 patients (55% male, mean ± SD age 37 ± 28 years) were obtained from the retrospective specimen collection. Fifty-nine isolates were obtained from clinical specimens and 33 from screening swabs. Among these isolates, 46 were obtained from skin and soft tissue samples and 13 from other body sites. A total of 59 isolates were obtained from the prospective CA-MRSA surveillance system from 59 patients (mean ± SD age 33 ± 21 years, male:female ratio 2.7).

Table 1 shows that most CA-MRSA strains isolated during 2002–2005 were associated with skin and soft tissue infections. Most cases of infection or colonization were associated with migration or travel history. Four healthcare workers acquired CA-MRSA strains epidemiologically unrelated to each other. In 2 instances, family members of these workers were also affected.

Strains were rarely resistant to clindamycin (5%), gentamicin (8% only in isolates recovered after 2002), or rifampicin (<1%). Susceptibility to cotrimoxazole was 89% during the first period and 100% during the second period. PVL-positive isolates remained multidrug susceptible throughout the study period, and were distinct from our endemic nosocomial strain.

Figure

Figure. Incidence of non–multidrug-resistant methicillin-resistant Staphylococcus aureus (MRSA) strains, Geneva, Switzerland, 1993–2005. A) Number of strains collected since 1993 showing an atypical multidrug-susceptible phenotype (white bars). Also shown are the number of...

The Figure, panel A, shows the incidence of isolates fulfilling our entry criteria and the proportion of strains producing PVL or harboring SCCmec IV or V. An increase in non–multidrug-resistant MRSA was observed during 1994–1997 (incidence 2.3 cases/10,000 admissions in 1997), and a second peak was observed during 2002–2005 (incidence 4.3 cases/10,000 admissions in 2004). Molecular characterization of the 151 strains showed that 124 (82%) harbored either SCCmec IV or V. A total of 92 isolates (61%) harbored at least 1 toxin gene, most frequently PVL (n = 60), followed by TSST-1 (n = 22) and exfoliatin A (n = 11). An isolate (ST149-MRSA-IV) from a Libyan patient harbored the PVL and TSST-1 genes (Table 2). A strain with PVL (ST80-MRSA-IV) was isolated in 1994 from a 73-year-old man from Libya. A case of bacteremia with PVL-positive CA-MRSA (ST80-MRSA-IV) was documented in a 28-year-old Tunisian woman who had an abscess of her left forearm in 2000. No case of necrotizing pneumonia was observed.

From 1994 through 1999, we identified 14 PVL-positive MRSA isolates. The Figure, panel B, shows that CA-MRSA identified during 1993–2002 consisted mainly of 3 clonotypes (ST80, ST88, and ST5). After 2002, these strains were less frequent and the proportion of other clonotypes increased.

Appendix Figure

Appendix Figure. Clustering trees. A) Trees obtained by using multiple-locus variable-number tandem repeat analysis (MLVA)–based genotyping for strains of Staphylococcus aureus harboring clinically important toxins, Geneva, Switzerland, 1993–2005. Year of isolation, staphylococcal cassette...

The Appendix Figure shows that toxin-harboring strains segregated in 18 multiple-locus variable-number tandem repeat analysis profiles and yielded 14 multilocus sequences types. ST80 (n = 39) was the most abundant type (42% of toxin-producing isolates). Other clusters contained well-described ST5 (TSST-1 or PVL positive), ST30, and ST8 (USA300) strains harboring the PVL gene.

Several epidemiologically linked cases were identified in the second period (Appendix Figure: a cluster of 6 family members with recurrent furunculosis over 5 years (ST80-MRSA-IV), 4 smaller family clusters with the same clone, and 3 family clusters with other clonally related strains (ST5-MRSA-IV, ST59-MRSA-V, ST1-MRSA-V). Two inmates incarcerated in the same cell of the Geneva prison had abscesses caused by the ST8-MRSA-IV (USA300) strain. One outbreak involved 5 neonates and 2 mothers colonized or infected with ST5-MRSA-IV harboring the PVL gene (5). We also observed a cluster of 5 patients from Kosovo infected or colonized with a PVL-producing strain (ST152-MRSA-V) resistant to gentamicin and amikacin.

Strains lacking toxins (Appendix Figure, panel B) showed a wide diversity of patterns. Most isolates (n = 33) were obtained from the first stain collection and showed many different genetic backgrounds (n = 21).

We studied 2 collections of non-multiresistant MRSA strains identified over a 13-year period at our institution. Our analysis showed that sporadic PVL-positive CA-MRSA has been isolated in Geneva since 1994; the largest cluster corresponded to ST 80 (SCCmec IV, PVL positive); the PVL gene is disseminated in many genetic backgrounds; strains showed diversity of genomic content; several epidemiologic clusters were identified; and many cases were linked to migration and travel.

Our data showed that resistance to fusidic acid or susceptibility to gentamicin should not be used as phenotypic criteria for CA-MRSA in Europe. For example, gentamicin-resistant ST152-MRSA-V found in 5 patients from Kosovo is common.

The number of genetic profiles identified appears particularly high, which is similar to profiles described in European countries, northern Africa, Oceania, and the Americas. In contrast to CA-MRSA epidemiology described as homogeneous in Australia (13), some parts of the United States (14), or Sweden (15), our local epidemiology appears more heterogeneous.

Frequently observed importation of CA-MRSA may enhance genetic exchange between strains and supports the need for active surveillance. Most CA-MRSA remained susceptible to many antimicrobial drugs, but genetic exchange between strains resulting in acquisition of resistance determinants in CA-MRSA or transfer of virulence markers into HA-MRSA are important concerns.

Our study has limitations. First, we used 2 strain collections. The collection obtained before 2003 may have omitted gentamicin-resistant CA-MRSA strains (e.g., ST152). Second, retrospective case ascertainment does not distinguish invasive from colonizing strains in all patients. Finally, we cannot exclude detection bias caused by our active MRSA screening policy (6,7).

In summary, increasing incidence of PVL-producing type IV CA-MRSA isolates is worrisome and indicates emergence of new MRSA lineages with a particular fitness for community transmission. Further epidemiologic and molecular typing studies are needed to document CA-MRSA carriage and infection rates and implement adequate infection control guidelines.