In vitro dilution models are used to simulate in vivo drug concentration-time profiles and thus to study the effects of various antibiotic concentrations on the bacteria investigated. The major disadvantage of these models is permanent dilution of the bacterial culture, which falsifies the resulting kill curves. Known equations, which usually correct bacterial loss by simple first-order kinetics, do not take into account special test conditions, such as variable elimination rate constants, exceptionally long periods of investigation, or formation of biofilms. In the present investigation, we examined the validity of these equations with regard to the test conditions mentioned. We simulated the concentration-time curves resulting from continuous infusion of 1,000 mg of meropenem with steady-state levels of 2.5, 5.0, and 7.5 g/ml in an in vitro dilution model. The resulting kill curves were compared with the kill curves obtained from incubation of bacteria in an undiluted system with meropenem at constant concentrations corresponding to the above-mentioned steady-state levels. Comparison of the matching kill curves showed that the common corrections, which do not consider the formation of biofilms in the compartments, partly overestimated the effect of bacterial dilution. We defined a factor, f, as an extension to the known equations which compensates for the effect of biofilms. Another extension was developed to allow the investigation of variable elimination rate constants. With the help of these extended mathematical corrections, we were able to fit the kill curves resulting from the in vitro dilution model exactly to the kill curves given by an undiluted system. Several in vitro kinetic models which simulate in vivo drug concentration-time profiles use continuous dilution to fit drug concentrations to the corresponding serum pharmacokinetics (4). The major disadvantage of these models is inevitable dilution of the bacterial culture, which falsifies the resulting kill curves. Several attempts have been made to compensate for this effect. (i) The antimicrobial activity of constant concentrations of antibiotics may be examined in a static system in which the bacterial culture is not diluted during the investigation (12). (ii) Dynamic in vitro models which do not allow bacteria to be washed out have been developed (1, 2). (iii) The loss of bacteria in an in vitro dilution model may be compensated for by mathematical correction.In the present report, we have summarized the previously existing equations (5,8,11) and investigated their validity with regard to different test conditions by simulating continuous infusion of meropenem with steady-state levels of 2.5, 5.0, and 7.5 g/ml in an in vitro dilution model. The resulting time-kill curves were compared with the kill curves resulting from incubation of test bacteria with meropenem at constant concentrations of 2.5 to 7.5 g/ml in a static system (batch culture). We assumed that bacterial growth in the batch culture is not limited by the supply of nutrients as long a...

In an in vitro dynamic model we compared the antimicrobial effects of two carbapenems, imipenem (MIC, 1 microg/ml) and meropenem (MIC, 0.25 microg/ml) on Pseudomonas aeruginosa. The antibiotics were administered either as short-time infusions once or three times a day or as continuous infusions with steady-state levels ranging from 0.5 to 20 microg/ml. From the resulting kill curves the period of time until the onset of bacterial death (dt), the rate constant of bacterial death (ka), the maximal reduction of CFU (mr), and the period of time until bacterial regrowth occurred (tr) were determined. Additionally, the occurrence of bacterial resistance during the simulations (rq) and the postantibiotic effect (PAE) were recorded. For both investigated carbapenems no significant difference in dt, ka, mr, and PAE values between the short-time infusions and continuous infusions with steady-state levels above 2 microg/ml could be detected. The tr was longest with continuous infusions of over approximately 24 h, corresponding to steady-state levels of 3 microg/ml for imipenem and 2.5 microg/ml for meropenem. An increase in MIC was observed only during continuous infusions with steady-state levels below 2 microg/ml. Independent of the chosen method of application and despite the lower MIC of meropenem, imipenem was slightly more effective than meropenem.