In-vitro studies of antibiotic combinations with special emphasis on the evaluation of newly developed methods

We have described an in-vitro pharmacokinetic model that mimics the serum and tissue concentrations of antibiotics during therapy of human patients, and thus presents a changing concentration of antibiotics to the bacterial inoculum. This pharmacokinetic model has been used to study antibiotic combinations that are used in the treatment of infections in granulocytopenic patients. In this model the addition of piperacillin to amikacin or of ceftazidime or azlocillin to another aminoglycoside prevented the regrowth of resistant subpopulations of Pseudomonas aeruginosa. The activities of several antibiotic combinations were studied in the in-vitro model, the conventional checker-board method and the time-kill method. Discrepant results were found with the model and the conventional tests in one-third of the combinations. The model may be useful in the preclinical study of antibiotic combinations and may be proved more predictive of clinical outcome than conventional tests.