Studies of in vitro and in vivo bactericidal interactions of vancomycin plus rifampin against Staphylococcus aureus have yielded conflicting results. In this study the efficacy of this drug combination in experimental endocarditis due to a methicillin-resistant strain of S. aureus was investigated. Left-sided endocarditis was induced in 84 rabbits by an infecting strain that had been found to be synergistically killed by vancomycin plus rifampin in vitro when tested by the timed-kill curve technique; in contrast, the checkerboard technique had indicated that the two drugs were antagonistic against this strain. Infected animals received no therapy, vancomycin alone (30 mg/kg per day), rifampin alone (20 mg/kg per day), or both drugs (in the same doses). The combination was significantly more effective than the single-drug regimens in terms of (1) reduction of mean methicillin-resistant S. aureus vegetation titers (P less than .05-.0005), (2) rate and incidence of sterilization of vegetations (P less than .0005), and (3) rate of "radical" cure of endocarditis (P less than .005). Vancomycin alone and vancomycin plus rifampin were equally effective in reducing mortality and sterilizing renal abscesses. The use of vancomycin prevented the in vivo development of resistance to rifampin. No evidence that rifampin exerted an antagonistic effect on the in vivo bactericidal activity of vancomycin was found.