In a prospective, randomized trial, teicoplanin (at a 400-mg intravenous loading dose followed by 200 mg/day intravenously or intramuscularly) was compared with flucloxacillin (8 g/day) in patients with severe staphylococcal infections. Teicoplanin proved unsatisfactory for the following reasons: failures or relapses were more frequent in the teicoplanin group, and blood levels were difficult to predict and tended to be low 24 hr after the loading dose. Future trials with this agent should use much-higher doses.