A dynamic in vitro model was used to assess four different vancomycin dosing regimens against Staphylococcus aureus. These regimens achieved peak drug concentrations of 48 ,ug/mI (single dose) and 30 ,ug/ml (dosed every 12 h) and constant concentrations of 16 and 8 ,ig/ml. Analysis of the area under the bacterial concentration-time curve, area under the first moment of the bacterial concentration-time curve, and bacterial elimination rate constant showed no difference in the rate or extent of bacterial killing. The optimal dosing method may be that which achieves the lowest area under the curve while concentrations are maintained above the MBC. (14)(15)(16)(17)20). Similarly, most in vitro experiments using fixed concentrations of vancomycin have not demonstrated concentrationdependent killing of various microorganisms (1,18,19). These studies suggest that the best dosing strategy would be to maintain the total area under the concentration-time curve (AUC) as small as possible to reduce dose-related toxicity while still maintaining adequate efficacy.In this study we assessed the efficacy of vancomycin against Staphylococcus aureus over a 24-h period in a dynamic in vitro model using a variety of dosing regimens, including some that are common in clinical practice. These experiments were designed to compare the efficacies of three different dosing schedules with the same AUC and a schedule with a smaller AUC.The dynamic in vitro model used in this experiment has been described previously (3)(4)(5) 37°C. At time zero a bolus of vancomycin (Sigma Laboratories; lot 20H0377) was added to each central chamber to achieve the desired initial concentration. A personal computer controlled two series of pumps with a purpose-developed computer program. The first series of pumps were syringe pumps which infused broth to dilute the contents of each central chamber at a predetermined rate, resulting in a desired exponential decline of vancomycin concentrations (with a half-life of 6 h). The second series were peristaltic pumps which returned the volume of the central chamber to its original level by removing the excess volume each hour. This excess provided a sample which was used to count the number of viable CFU per milliliter. The MIC was estimated prior to the study and following exposure to vancomycin and was 0.75 ,ug/ml by the standard dilution tube technique, with an inoculum size of 2.3 x 105 CFU/ml. The test strain used for the experiments was incubated in MHB overnight and diluted 2.5 h before the experiment to produce an inoculum of _107 CFU/ml in log-phase growth.The study assessed the following vancomycin regimens: (i) a peak concentration of 48 ,ug/ml and a trough of 3 ,ug/ml at 24 h (AUC from 0 to 24 h [AUCO24] = 389 ,ug -h -ml-'); (ii) a peak concentration of 30 ,ug/ml and a trough of 7.5 ,ug/ml at 12 h, dose repeated at 12 h (AUCO24 = 389 ,ug* h * ml-1); (iii) a constant concentration of 16.2 ,ug/ml, representing a 24-h infusion (AUCO024 = 389 ,ug h ml-1); and (iv) a constant concentration of 8 ,ug/ml, repres...