Background: Methicillin-resistant Staphylococcus aureus (MRSA) infection typically occurs in chronically ill patients requiring long-term antimicrobial therapy or hospitalization. However, community-associated MRSA (CA-MRSA) necrotizing soft tissue infections seem to be increasing in incidence. Our aim was to describe the incidence and microbiologic characteristics of CA-MRSA isolates collected at an army community hospital. Methods: We report a retrospective review of MRSA isolates identified during 1998-2003 at the microbiology laboratory of Moncrief Army Community Hospital that serves a community of approximately 40,000 transient residents yearly in Fort Jackson, South Carolina. We evaluated the incidence of MRSA in our laboratory during 1998-2003. For MRSA isolates from 2003, we evaluated antimicrobial susceptibility patterns. Six selected isolates were evaluated by molecular typing, resistance gene analysis, and toxin analysis. Results: During 1998-2003, 241 (23%) of 1041 S. aureus isolates identified at the hospital microbiology laboratory were resistant to methicillin. Of these 241 MRSA isolates, 223 were cultured from outpatients. The incidence of MRSA in our population increased from 12% of S. aureus isolates in 1998 to 43% in 2003. In 2003, MRSA was cultured from 76 different patients. Isolates of MRSA were often resistant to erythromycin (91%), although resistance to other agents was less common: Ciprofloxacin (14%), levofloxacin (14%), clindamycin (3%), tetracycline (3%), and trimethoprim sulfamethoxazole (1%). No isolates were resistant to vancomycin, gentamicin, nitrofurantoin, or rifampin. Six CA-MRSA isolates were compared by pulsed-field gel electrophoresis (PFGE). Five were PFGE type USA300, and one was PFGE type USA100, based on the U.S. Centers for Disease Control and Prevention (CDC) classification scheme. The five USA300 isolates carried SCCmec type IV, and the USA100 carried SCCmec II. None of the isolates were positive by PCR for genes encoding enterotoxins A-E and H, or toxic shock syndrome toxin (TSST-1), but the five USA300 isolates carried the gene coding for Panton-Valentine leukocidin toxin. Conclusions: The incidence of MRSA at our institution is increasing. Isolates of MRSA show resistance patterns and microbiologic characteristics consistent with CA-MRSA isolates from the United States. Clinicians should consider the possibility of CA-MRSA in patients with soft-tissue infections who do not respond to initial therapy with beta-lactam antimicrobial agents.