New concept and a theoretical consideration of the mechanism-based pharmacokinetics/ pharmacodynamics (PK/PD) modeling for antimicrobial agents.

Pharmacodynamic (PD) characterization (concentration-dependent, time-dependent, etc.) of antibiotics is determined by aspects of the pharmacodynamic interaction between antibiotics and microorganisms. There are three major aspects of the pharmacodynamic interaction between antibiotics and microorganisms; 1) the minimum drug concentration required for the exhibition of antibacterial activity (MIC: minimum inhibitory concentration, MBC: minimum bactericidal concentration, etc.), 2) the relationship between drug concentration and bactericidal activity and 3) the magnitude of any persistent antibiotic activity (sub-MIC effect, post antibiotic effect, etc.). In the PK/PD approach based on the MIC (static MIC approach), information concerning aspect 1) alone is treated as the quantitative PD parameter (MIC), while information concerning aspects 2) and 3) are not represented as quantified PD parameters in spite of their importance in in vivo pharmacodynamic situation. On the other hand, in the PK/PD approach based on the time-kill profile (dynamic PK/PD approach), information concerning aspects 1) - 3) can all be represented as quantitative dynamic PD parameters (epsilon; the maximum kill rate constant, gamma; the Hill coefficient and EC50; the antibiotic concentration at which 50% of the maximum effect is obtained) together with the growth rates of the organisms (lambda). We thought that the PD characterization of antibiotics should be determined by integrating the dynamic PD parameters and the growth rates, so we developed a new concept integrating these parameters so that a good approximation of the time course of in vivo antibacterial activity exhibited by a antibiotic might be predicted from these parameters and the pharmacokinetics of the drug. To achieve this, we analyzed the time-kill profiles of a wide range of antibiotics against various microorganisms and obtained the dynamic PD parameters and the growth rates for the various combinations of antibiotics and microorganisms. Then we analyzed the causal relationship between the PD characteristics of the antibiotics and the dynamic PD parameters and the growth rates. As a result, we derived the following criteria for predicting the PD characteristics of antibiotics. (i) if the epsilon/lambda is greater than about 10 and gamma is less than one, the pharmacodynamics should be concentration-dependent (ii) if the epsilon/lambda is within the range 1-2 and gamma is about 5 or more, the pharmacodynamics should be time-dependent (iii) if the epsilon/lambda is within the range 1-4 and gamma is in the range 1-12, the pharmacodynamics should be time-dependent or both time- and concentration-dependent. The PD characterization of antibiotics against various strains of different microorganisms can be predicted relatively easily and quickly by utilizing these criteria. These findings make it possible to determine the kinds of causative pathogens to which the antibiotics should be applicable in the clinical sites. Furthermore, it is expected that effective strategies for development and establishing the optimum dosage regimen of novel antibiotics could be worked out more scientifically and efficiently than ever on the basis of the PK/PD parameters corresponding to the predicted PD characters.