Sir, Historically, methicillin-resistant Staphylococcus aureus (MRSA) has been regarded as a nosocomial pathogen responsible for severe toxin-mediated disease or invasive pyogenic infections. In recent years, however, community-associated MRSA (CA-MRSA) has been reported from around the world. Although the first CA-MRSA strains were susceptible to most antimicrobials, antimicrobial resistance has increased. One CA-MRSA clone, identified by the CDC as pulsotype USA300, has been implicated in outbreaks within the USA and is resistant to many currently marketed antimicrobial agents. This clone carries both the SCCmec IV element and the Panton–Valentine leukocidin (PVL) locus. The aim of this study was to compare the activity of a novel ketolide, cethromycin (previously ABT-773), with potentially improved pharmacodynamic and therapeutic benefits, three other macrolides (azithromycin, clarithromycin and erythromycin) and one lincosamide (clindamycin) against the 170 MRSA USA300 isolates that are in the Center for Biological Defense (CBD) collection. Seventeen MRSA isolates were procured from two hospital laboratories in Seattle, WA, USA. These were isolated from cultures of blood, skin lesions and tissue abscesses of patients seen in their clinics or emergency department in 2003 and 2004. Of the remaining 153 MRSA, 121 isolates were obtained from cultures of lesions, abscesses and nasopharyngeal specimens from outpatients seen at west central Florida clinical laboratories and doctor’s offices, whereas 32 MRSA were isolated from lesions, abscesses, nasopharyngeal and blood specimens acquired from patients seen at hospitals in the Tampa Bay area from 2004 to 2006. These isolates were previously characterized as USA300 by PFGE using Network on Antimicrobial Resistance in Staphylococcus aureus (NARSA) standards for comparison. Cethromycin (obtained from Advanced Life Sciences, Woodridge, IL, USA) was prepared and diluted with calciumadjusted Mueller–Hinton broth (Remel, Lenexa, KS, USA) using 96-well microtitre plates following standard protocols. The MIC and MBC were determined using orthodox procedures. Clarithromycin, erythromycin and clindamycin were tested with Sensititre by TREK Diagnostic Systems (Cleveland, OH, USA), whereas azithromycin, erythromycin and clindamycin were tested using Etest (AB Biodisk, North America Inc., Piscataway, NJ, USA), according to manufacturer’s instructions. All MIC interpretations followed CLSI guidelines. Clindamycin resistance was examined further with the D-test. Controls were S. aureus ATCC 29213, S. aureus ATCC BAA977 and S. aureus ATCC BAA976. Cethromycin was effective against all 170 isolates (Table 1). The cethromycin MIC range was 0.002–0.125 mg/L, whereas the MIC50 and MIC90 were both 0.002 mg/L. The cethromycin MBC range was 0.002–1.0 mg/L, with the MBC50 of 0.002 mg/L and the MBC90 of 0.008 mg/L. Of the 170 CA-MRSA, 169 (99.4%) were intermediate or resistant to at least one macrolide. In contrast, 164 (96.5%) isolates were susceptible to clindamycin. Only six (3.5%) isolates demonstrated either constitutive resistance [4 (2.4%)] or inducible resistance [2 (1.2%)] to clindamycin. This is unlike the USA300 isolates in Boston where 57% of the isolates were clindamycin-resistant. Against 10 erythromycin-susceptible isolates, the cethromycin MIC and MBC ranges were 0.002–0.008 mg/L, compared with the MIC range of 0.002–0.25 mg/L and MBC range of 0.002–1.0 mg/L obtained with the erythromycin-resistant isolates. However, MIC50 and MBC50 values, as well as MIC90 and MBC90 values, did not change regardless of erythromycin susceptibility. There were six isolates resistant to both erythromycin and clindamycin, against which cethromycin MIC and MBC ranges were 0.002–0.06 mg/L. This population’s susceptibility to cethromycin is especially important in the light of the fact that most isolates were resistant to one or more macrolides. This collection of MRSA USA300 isolates was overwhelmingly macrolide-resistant and clindamycinsusceptible. The study of Chavez-Bueno et al. demonstrated a similar pattern in CA-MRSA isolates among children in Texas where the inducible resistance to clindamycin was slowly replaced by clindamycin susceptibility (non-inducible resistance). Most of the CA-MRSA isolates in the Texas study carried the erm(C) gene, with others carried erm(A) or msr(A) with or without erm(C). This phenotypic change suggests an alteration in the proportion of the macrolide resistance determinants in the population. Tenover et al. reported the msr(A) gene in isolates comprising a specific clone (USA300-0114). Accordingly, we hypothesized that most of the CBD collection of MRSA USA300 isolates (only one being USA300-0114) also carried a different determinant, putatively the msr(A)/msr(B) gene encoding an efflux protein that normally renders a bacterial cell resistant only to macrolides and streptogramin B compounds (MS phenotype) instead of the MLSB resistance genes [i.e. erm(A) and erm(C)]. Using the oligonucleotides described by Sutcliffe et al. for PCR primers and a well-characterized MRSA strain NARSA 384 (USA300-0114) as positive control, we performed a PCR assay for the msr(A) gene on DNA extracted from all 170 isolates. After the amplicons were electrophoresed on a 1% agarose gel, a positive PCR reaction yielded a DNA band of 399 bp. All negative and questionable results were performed three times. The assay produced a positive reaction in 160 (94.1%) isolates, whereas only 10 (5.9%) of the 170 MRSA yielded negative PCR reactions. These Research letters