We enrolled 14 human volunteers in a crossoverex vivostudy comparing 6 regimens including amikacin (0.5g), ceftazidime (2g), cefoperazone (2 and 4g) and the combination of each cephalosporin (2g) with amikacin. The antibiotics were given by intravenous infusion over 30 minutes. Blood samples were obtained 1, 4, 6 and 12 hours after the end of infusion to measure serum concentrations, bactericidal titres and killing rates. Two parameters were derived from the killing curves: the initial rate of killing and the relative bioactivity (area under the time-kill curve as a percentage of the control growth curve). The elimination half-life of cefoperazone was decreased by concomitant administration of amikacin, resulting in higher trough levels. Although synergy was seen in the checkerboard method in half of the strains, serum bactericidal titres unusually showed synergy which was more frequent with cefoperazone + amikacin than with ceftazidime + amikacin. For both cephalosporins, the relative bioactivity, but not the killing rate, was dependent on concentration. Ceftazidime had a higher intrinsic activity despite the higher serum concentrations obtained with cefoperazone. The addition of amikacin conferred concentration dependency of the killing rate to the combination. None of the regimens tested provided enough antimicrobial activity to support once daily administration.