An open study was carried out on 16 patients with hospital-acquired, bacteraemic Staphylococcus aureus infections to evaluate the safety and efficacy of teicoplanin plus rifampicin. Patients received teicoplanin 400 mg bd for the first 24 h followed by 400 mg od thereafter, and rifampicin 600 mg bd. Both agents were given intravenously. Serum samples were collected to determine trough and peak antibiotic concentrations. The MIC of teicoplanin and rifampicin and the MBC of teicoplanin were determined for all S. aureus isolates. Time-kill curves were performed for the drugs individually and in combination. Clinical efficacy was assessed by the APACHE II scoring system. Bacteriological success was evaluated by elimination, persistence or recurrence of S. aureus. Safety was carefully monitored by regular biochemical and haematological testing and recording of adverse events. Fifteen patients were evaluable, of whom 13 (86.7%) were clinically cured with elimination of S. aureus. One patient died, but death was not attributed to the study drugs. Treatment failed in another patient who relapsed with a high fever. S. aureus was recovered from blood cultures from this patient, and resistance to rifampicin had developed. Time-kill curves all showed adequate killing of S. aureus at the drug concentrations measured in vivo. Neither synergy nor antagonism between teicoplanin and rifampicin was demonstrated. The combination of teicoplanin and rifampicin is an effective and well-tolerated treatment for bacteraemic S. aureus infections, but in deep-seated foci of infection resistance to rifampicin may develop.