Abstract The changing epidemiology of Staphylococcus aureus bacteremia has been noted worldwide. This enhanced awareness appears to be closely associated with the evolution of health care systems. To further delineate this change and to clarify the prevalence of true community-acquired methicillin-resistant S. aureus (MRSA), reclassification for community-onset bacteremia was proposed. Exposure to health care system, such as nursing home residence, regular outpatient invasive interventions, and prior hospitalization within 1 year, was identified among the community-onset S. aureus bacteremia patients. During the 1-year study period, 102 episodes of S. aureus bacteremia from the emergency department patients of a teaching hospital were prospectively enrolled. Nine of the episodes were hospital-acquired, 56 episodes were associated with health care system exposure, and the remaining 37 episodes were classified as true community-acquired bacteremia. The characteristics of patients, primary site of infection, antimicrobial susceptibilities of S. aureus isolates, adequacy of initial antimicrobial therapy, and percentage of metastatic infections differed significantly between health care-associated and true community-acquired S. aureus bacteremias. Prevalence of MRSA infection in true community-acquired bacteremia was low in contrast to bacteremia with health care-associated exposure (2.7% versus 42.9%, P < 0.01). In conclusion, clinical characteristics and risk of contracting methicillin-resistant S. aureus bacteremia among community patients with and without exposure to health care system are distinct. Precise classification of patients is mandatory for the surveillance of antimicrobial resistance and selection of rationale empirical antibiotics.