It is important when using PD data to remember the four factors necessary for making a rational decision on drug dosage and comparison between agents. These include looking at the PD target, looking at the population variability by using population PK modeling, identifying the target organisms and their MIC distributions, and then correcting for protein binding.

To further develop optimal ways of using PD data in clinical decision-making, we carried out a number of studies. Using an AUC/MIC target ratio of 125, which was demonstrated in the past to be effective when managing P. aeruginosa, we evaluated the target attainment according to MIC for levofloxacin 750 mg o.d. Plasma target attainment was 100% at MIC values of 0.25 mg/ml, 0.5 mg/ml, and 0.75 mg/ml, but at 1.0 mg/ml, this drops dramatically to 54%, and at 2.0 mg/ml, it drops to 0%. However, while ELF target attainment dropped rapidly from 100% to 80% at 0.5 mg/ml, the drop was then much gentler than that for plasma, with 60% target attainment at 1.0 mg/ml, 40% at 2.0 mg/ml, 20% at 4.0 mg/ml, and > 10% of patients attain the target even for pathogens with a higher MIC. The explanation for why some patients with pathogens with low MIC values fail microbiologically, whereas other patients with pathogens with much higher MIC values do well, may be that in the latter group there is better ELF penetration.