In vitro models for the study of combination antibiotic therapy in neutropenic patients.

Neutropenic patients are at risk of serious infection caused by gram-negative bacilli and staphylococci. The mortality rate associated with gram-negative bacteremia in these patients is extremely high, especially in those with persistent and profound granulocytopenia. In these latter patients, the best results have been obtained by administering combinations of antibiotics in which both agents are active and/or show in vitro synergism against the infecting organism. Most combinations include an aminoglycoside such as amikacin and a broad-spectrum beta-lactam antibiotic, such as azlocillin, mezlocillin, piperacillin, or ceftazidime. The International Antimicrobial Therapy Project Group of the European Organization for Research and Treatment of Cancer has completed several studies evaluating various antibiotic combinations in the empiric treatment of febrile neutropenic patients. These trials have evaluated cephalothin plus gentamicin, carbenicillin plus gentamicin, and cephalothin plus carbenicillin; carbenicillin plus amikacin and carbenicillin plus amikacin plus cefazolin; azlocillin plus amikacin, ticarcillin plus amikacin, and cefotaxime plus amikacin; and azlocillin plus amikacin versus ceftazidime plus long- or short-course amikacin. The preclinical evaluation of antibiotic combinations usually involves the in vitro testing of antibiotics alone and in combination by the checkerboard method or with the use of time-kill curves. However, these methods expose the bacterial culture to a static or constant concentration of the drugs. During the in vivo treatment of infections, bacteria are exposed to changing concentrations of antibiotics, which are contingent on the individual pharmacokinetics of these drugs. We have designed a two-compartment in vitro pharmacokinetic model that allows the simultaneous study of the activity of two antibiotics with similar or different half-lives against a number of bacteria. Amikacin and azlocillin have been studied alone and in combination in this model against Pseudomonas aeruginosa, a frequent cause of bacteremia in neutropenic patients. In pharmacologically relevant doses, amikacin alone produced rapid bacterial killing, followed by regrowth of resistant subpopulations. Azlocillin alone produced a more gradual reduction of the bacterial inoculum, with ultimate bacteriostasis. Amikacin plus azlocillin produced rapid and complete eradication of the organism. In vitro pharmacokinetic models may prove to be more predictive of clinical outcome than are traditional static in vitro methods used to study antibiotic combinations.