Antifungal pharmacokinetics and pharmacodynamics

Abstract

Successful treatment of infectious diseases requires choice of the most suitable antimicrobial agent, comprising consideration of drug pharmacokinetics (PK), including penetration into infection site, pathogen susceptibility, optimal route of drug administration, drug dose, frequency of administration, duration of therapy, and drug toxicity. Antimicrobial pharmacokinetic/pharmacodynamic (PK/PD) studies consider these variables and have been useful in drug development, optimizing dosing regimens, determining susceptibility breakpoints, and limiting toxicity of antifungal therapy. Here the concepts of antifungal PK/PD studies are reviewed, with emphasis on methodology and application. The initial sections of this review focus on principles and methodology. Then the pharmacodynamics of each major antifungal drug class (polyenes, flucytosine, azoles, and echinocandins) is discussed. Finally, the review discusses novel areas of pharmacodynamic investigation in the study and application of combination therapy.

Figure 1.

Pharmacokinetic/pharmacodynamic relationship of antimicrobial dosing over time relative to organism MIC. The three PD indices are also listed, including C_max/MIC, AUC/MIC, and %T > MIC.

Figure 2.

The postantifungal effect. Shown in the figure is a representative example of a drug with a prolonged PAFE. Represented by solid circles is the growth of organism from the thighs of untreated infected mice (control). In comparison, shown by open triangles is the growth curve of organism from thighs of infected mice treated with a single 20-mg/kg dose of drug given at time point 0 h. The solid bar represents the time that drug concentrations are expected to exceed the MIC. The open circles represent the time at which drug concentration decreases below the MIC until the point at which 1-log₁₀ growth was observed. This time period (38 h) minus the time period it takes the control group to grow 1 log₁₀ (in this case 2.5 h) is the PAFE (i.e., ∼35.5 h).

Figure 3.

Concentration–time profile showing the effect of fractionating a total daily dose into once-, twice-, four-times-, and eight-times-daily fractions. AUC will remain the same because the total daily dose administered is the same in all four regimens; however, C_max will progressively decline and %T > MIC will progressively increase as the dose is fractionated into increasing fractions.

Figure 4.

Concentration–effect curves for a dose-fractionation experiment that included four fractionations. The total dose was kept the same for each regimen and is represented on the x-axis. Treatment outcome is represented on the y-axis as the change in fungal burden from the start of therapy. The dashed line represents net stasis over the duration (24 h). In this experiment, each of the fractionations essentially overlaps one another, indicating that AUC/MIC, as this is kept constant given that the same total dose is administered, is likely the PD index predictive of efficacy. If C_max was predictive, one would expect the q24h regimen to have enhanced effect (i.e., the concentration–effect curve would be significantly lower on the y-axis) compared with more fractionated regimens. However, if %T > MIC was the predictive index, one would expect the q3h regimen to have an enhanced effect (i.e., the concentration-effect curve would be significantly lower on the y-axis) compared with less fractionated regimens.

Figure 5.

Dose–response curves for each of the three PD indices (displayed on the x-axis as AUC/MIC, C_max/MIC, and time above MIC) in a dose-fractionation experiment. Shown on the y-axis is the change in organism burden from the start of the experiment, with the dashed line representing net stasis. Points above the dashed line represent an increase in organism burden (i.e., net growth), whereas those below the line represent a decrease in organism burden (i.e., net cidal activity). The curved line represents the best-fit line based on nonlinear regression modeling using the Hill equation (sigmoidal dose-response model) and in the legend is the coefficient of determination (R²). In the above example, the PD index that best predicts efficacy is AUC/MIC.