We administered 1 g of imipenem along with equal amounts of cilastatin (a dehydropeptidase I inhibitor) or 2 g of moxalactam intravenously over a period of 30 min to six volunteers in a crossover manner 1 week apart. The antibiotic concentrations and pharmacokinetics for each drug were determined and integrated with the microbiologic activity by measuring the duration of time that the free drug concentrations remained above the MICs for 90% of 581 clinical isolates and by measuring serum bactericidal activities against organisms which commonly infect granulocytopenic cancer patients. Moxalactam produced serum levels at 1 h after infusion of 99.9 micrograms/ml; these levels were four times greater than the plasma levels of imipenem (22.8 micrograms/ml). The trough (5.5-h) moxalactam serum levels were 10 times greater than those of imipenem (18.5 and 1.7 micrograms/ml, respectively). Essentially all of the imipenem was unbound to protein, whereas 36 to 42% of the moxalactam was unbound. Moxalactam produced free antibiotic concentrations that were above the MIC for 90% of the strains tested for more than 6 h against all of the species tested except Staphylococcus aureus (5.3 h), Enterobacter hafnia (1.6 h), and Pseudomonas aeruginosa (0 h). The imipenem concentrations were above the MIC for 90% of the strains tested for 5.6 h or more against all of the bacteria tested except Proteus spp. and Pseudomonas aeruginosa (4.5 h). The geometric mean peak bactericidal titers from volunteers receiving imipenem were more than 1:8 against all bacteria and were significantly higher than the titers from volunteers receiving moxalactam against S. aureus (1:7.3) and Pseudomonas aeruginosa (1:4.5). These data, in addition to information obtained from animal models, indicate that imipenem is a promising new candidate for carefully controlled clinical trials as a single agent for therapy of serious infections, including empiric therapy for fever in granulocytopenic cancer patients.