Bacteria growing in vivo multiply much more slowly than in vitro. If the bactericidal activity of quinolones may be affected by an increase in generation time (g) was studied in batch cultures as well as under the well-controlled conditions of a continuous-flow culture. By limiting the nutrient supply, generation times were lengthened from approximately 0.45 to 1.5 h up to 3.9 h. Three recent clinical isolates each of Escherichia coli, Staphylococcus aureus and Pseudomonas aeruginosa were exposed to twice the MIC of ciprofloxacin, norfloxacin, fleroxacin and ofloxacin. The 'killing rates' were calculated in analogy to the growth rate. The bactericidal activity of the quinolones tested against E. coli was minimally influenced by the reduced generation time and the effect against S. aureus was moderate. As compared to their rapidly growing counterparts (g = 0.4 h) slowly growing (g = 1.3 h) P. aeruginosa were killed even more effectively by ciprofloxacin (176% increase) fleroxacin (48% increase) norfloxacin (36% increase) and ofloxacin (86% increase). These changes may likely be due to adaptive responses of the outer membranes of the bacteria to the limited nutrient supply thereby sensitizing the bacteria to the bactericidal activity of quinolones.

On the basis of a review of published literature it is demonstrated that pharmacokinetic parameters of antibacterial agents correlate well with therapeutic efficacy in animal models, provided pharmacodynamic parameters are taken into account. The time that serum levels exceed the MIC is the most significant parameter determining efficacy of β-lactams, whereas the efficacy of aminoglycosides is dependent on serum concentrations and the area under the curve. The efficacy of quinolones tends to be correlated to the doses administered or drug levels achieved. However, specific pharmacodynamic properties contribute significantly to the therapeutic efficacy of a few quinolones only whereas other quinolones lack these specific pharmacodynamic attributes. Thus, the correlation of pharmacokinetic parameters with therapeutic efficacy provides important basic concepts for the design of preclinical and clinical studies in the course of which additional pharmacodynamic properties will become apparent.

Ten patients about to undergo a colorectal operation lasting an average of three hours received 500 mg each of imipenem and cilastatin i.v. preoperatively. During the operation blood and tissue samples were taken in order to determine the serum kinetics of the substances as well as the levels of imipenem in the cutis, subcutis, fascia, muscle, parietal peritoneum and colon. The imipenem concentrations were measured by HPLC. The mean peak serum concentration was 26 mg/l, the mean half-life 55 min and the AUC 40 mg/l/h−1. The serum pharmacokinetics of imipenem was subject to substantially larger fluctuations in this patient group than in subjects or patients without surgery. Imipenem rapidly penetrates into tissue, with peak concentrations being reached after 10–25 min. The highest imipenem concentrations were found in the colon, the lowest in the cutis and subcutis. After 1 h a level of 8 mg/kg imipenem was still found in the colon. The concentrations were > 1 mg/kg in all tissues for more than 3 h p. a. and thus within this period exceeded the MICs ofEscherichia coli andBacteriodes fragilis, the indicator organisms of intraabdominal infections.