Abstract In empirical antibacterial therapy, regional surveillance is expected to yield important information for the determination of the class and dosage regimen of antibacterial agents to be used when dealing with infections with organisms such as Pseudomonas aeruginosa, in which strains resistant to antibacterial agents have been increasing. The minimal inhibitory concentrations (MICs) of five carbapenem antibiotics against P. aeruginosa strains isolated in the Northern Kyushu district of Japan between 2005 and 2006 were measured, and 100 strains for which carbapenem MICs were ≤0.5–32 μg/ml were selected. In this study, MIC was measured by two methods, i.e., the common serial twofold dilution method and an integrated concentration method, in which the concentration was changed, in increments of 2 μg/ml, from 2 to 16 μg/ml. The MIC50/MIC90 values for imipenem, meropenem, biapenem, doripenem, and panipenem, respectively, with the former method were 8/16, 4/16, 4/16, 2/8, and 16/16 μg/ml; and the values were 6/10, 4/12, 4/10, 2/6, and 10/16 μg/ml with the latter method. The MIC data obtained with both methods were subjected to pharmacokinetic/pharmacodynamic (PK/PD) analysis with Monte Carlo simulation to calculate the probability of achieving the target of time above MIC (T>MIC) with each carbapenem. The probability of achieving 25% time above the MIC (T>MIC; % of T>MIC for dosing intervals) and 40% T>MIC against P. aeruginosa with any dosage regimen was higher with doripenem than with any other carbapenem tested. When the two sets of MIC data were subjected to PK/PD analysis, the difference between the two methods in the probability of achieving each % T>MIC was small, thus endorsing the validity of the serial twofold dilution method. [ABSTRACT FROM AUTHOR]