Staphylococcus aureus is the most common cause of hospital-acquired bacteremia. From 1995 through 2000, data on age, sex, patient specialty at time of first bacteremia, primary and secondary sites of infection, delay in initiating antimicrobial therapy, and patient outcome were prospectively recorded for 815 patients with nosocomial S. aureus bacteremia. The proportion of patients whose death was attributable to methicillin-resistant S. aureus (MRSA) was significantly higher than that for methicillin-susceptible S. aureus (MSSA) (11.8% vs. 5.1%; odds ratio [OR], 2.49; 95% confidence interval [CI], 1.46–4.24; P < .001). After adjustment for host variables, the OR decreased to 1.72 (95% CI, 0.92–3.20; P = .09). There was no significant difference between rates of disseminated infection (7.1% vs. 6.2% for MRSA-infected patients and MSSA-infected patients, respectively; P = .63), though the rate of death due to disseminated infection was significantly higher than death due to uncomplicated infection (37% vs. 10% for MRSA-infected patients [P < .001] and 37% vs. 3% for MSSA-infected patients [P < .001]). There was a strong statistical trend toward death due to nosocomial MRSA infection and bacteremia, compared with MSSA.