A New Marker of Echinocandin Activity in an In Vitro Pharmacokinetic/Pharmacodynamic Model Correlates with an Animal Model of Aspergillus fumigatus Infection

ABSTRACT

The lack of a quantifiable marker for echinocandin activity hinders in vitro pharmacokinetic/pharmacodynamic (PK/PD) studies for Aspergillus spp. We developed an in vitro PK/PD model simulating the pharmacokinetics of anidulafungin and assessing its pharmacodynamics against Aspergillus fumigatus with a new, easily quantifiable, sensitive, and reproducible marker. Two clinical A. fumigatus isolates previously used in animals (AZN8196 and V52-35) with identical anidulafungin EUCAST (0.03 μg/ml) and CLSI (0.015 μg/ml) minimal effective concentrations (MEC) and one isolate (strain AFU79728) with an MEC of >16 μg/ml were tested in a two-compartment PK/PD dialysis/diffusion closed model containing a dialysis membrane (DM) tube inoculated with 10³ CFU/ml. During anidulafungin exposure, two types of fungal forms were observed inside the DM tube: floating conidia that were quantified by cultures and aberrant mycelia that were quantified by the vertical height of the mycelia attached on the DM tube. No aberrant mycelia were found for the resistant isolate or in the drug-free controls. An in vitro exposure-effect relationship was similar to that found in animals using survival as an endpoint, with a free-drug area under the concentration-time curve from 0 to 24 h (fAUC_0–24) associated with 50% of maximal activity of 2.21 (range, 1.81 to 2.71) mg · h/liter in vitro versus 2.62 (range, 1.88 to 3.65) mg · h/liter in vivo (P = 0.41). The hillslopes were also similar, with 1.96 versus 1.34 (P = 0.29). Analysis of each isolate separately showed increased antifungal susceptibility between AZN8196 and V52-35 (P < 0.001) even though they have the same CLSI and EUCAST MECs, but the strains have two 2-fold dilutions lower MICs using Etest and the XTT {2,3-bis (2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino) carbonyl]-2H-tetrazolium hydroxide} method. Dose fractionation studies with all three echinocandins showed that their activities are best described by fAUC and not the maximum concentration of free drug (fC_max). The new marker correlated with in vivo outcome and can be used for in vitro PK/PD studies exploring the pharmacodynamics of echinocandins against Aspergillus spp.

INTRODUCTION

Echinocandins demonstrate in vitro activity against Candida and Aspergillus species (1). Although echinocandins have been extensively used for the treatment of invasive candidiasis, there are few data for the treatment of invasive aspergillosis, where echinocandins have been used mainly as salvage and/or combination therapy. In light of the emergence of azole-resistant Aspergillus fumigatus isolates, there is a need for more chemotherapeutic options for the treatment of invasive aspergillosis.

Echinocandins have a unique mode of action against Aspergillus spp., resulting only in morphological alterations with formation of short, stubby, and highly branched hyphal clusters without complete inhibition of growth. The minimal effective concentration (ΜΕC) is commonly used to determine the in vitro activity of echinocandins in static models. However, the MEC is subjective, technically troublesome, and not feasible for measuring antifungal activity over time and drug concentrations, and the MEC has no clear in vivo correlation. Other markers like galactomannan and fungal DNA have limited value for assessing antifungal activity of echinocandins because of the lack of a clear correlation with fungal growth (2, 3). Due to these limitations in reliably assessing fungal growth after exposure to echinocandins and thereby their antifungal effects, there are no in vitro pharmacokinetic/pharmacodynamic (PK/PD) models for echinocandins and Aspergillus spp. Furthermore, there are no in vitro models whose results are correlated with in vivo outcome.

In vivo PK/PD studies are limited, and efficacy is usually assessed using survival or PCR results as surrogate markers. However, conflicting results in terms of the pharmacodynamic index that best describes the activity of echinocandins against Aspergillus spp. were found. One study supported the maximum concentration of drug in plasma (C_max)/MEC ratio (4), and the other indicated the area under the concentration-time curve (AUC)/MEC ratio (5) as the driving PK/PD index for caspofungin and anidulafungin, respectively. Since echinocandins possess immunomodulatory properties, in vivo pharmacodynamics may be masked by those effects (6). Furthermore, animal toxicity prohibits the use of maximal doses that could be effective in human infections (5).

Given these shortcomings in determining the activity and pharmacodynamics of echinocandins against A. fumigatus, we developed a new in vitro marker that is easily quantifiable, reproducible, and amenable for PK/PD studies. We further used an in vitro PK/PD model that simulates anidulafungin pharmacokinetics and assessed its pharmacodynamics against A. fumigatus utilizing the new in vitro marker. This marker was observed for all three echinocandins and correlated with survival of mice with disseminated aspergillosis that were treated with anidulafungin.

RESULTS

Antifungal susceptibility.

Even though the EUCAST and CLSI MECs of anidulafungin are the same, the Etest and XTT {2,3-bis (2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino) carbonyl]-2H-tetrazolium hydroxide} method revealed two 2-fold dilution differences in in vitro anidulafungin susceptibility between the azole-susceptible AZN8196 and the azole-resistant V52-35 isolates. The Etest MICs were 0.023 μg/ml for AZN8196 and 0.094 μg/ml for the V52-35 after 24 h and 48 h. The XTT MICs were 0.002 and 0.008 μg/ml for AZN8196 and V52-35, respectively, with 22 to 27% higher metabolism observed for the V52-35 at sub-MIC concentrations of 0.002 and 0.004 μg/ml. The Etest and XTT MICs of the anidulafungin-resistant strain AFU79728 were >16 μg/ml.

In vitro pharmacodynamics.

The in vitro PK parameters free-drug C_max (fC_max) and half-life (t_1/2) were within the target values (<15% deviation). The percentage of floating conidia inside the dialysis membrane (DM) tube progressively decreased at higher anidulafungin exposures (Fig. 1). The median (range) coefficient of variation of the percentage of floating conidia among the replicates was 14% (1 to 71%) across all time points and for both isolates. The largest decline in the numbers of CFU was observed within the first 24 h. Within this period, the first aberrant mycelia attached on the membrane of the DM tube appeared, but they were more clearly visible after 48 h; their height did not change significantly over time. The percentage of aberrant mycelia increased at higher exposures of anidulafungin (Fig. 2). This new marker of antifungal activity was quite reproducible with coefficients of variation (CVs) of <10% among replicates. Interestingly, the sum of the percentage of floating conidia and the percentage of aberrant mycelia calculated as described in Materials and Methods was close to 100% (98% [77 to 106%] for AZN8196 and 97% [80 to 114%] for V52-35). This indicates that the conidia of the initial inoculum are either floating inside the DM tube without germinating or they are attached on the DM tube, forming the aberrant mycelia. No attached mycelia on the DM tube were observed for the drug-free control and the resistant AFU79728 strain even at the highest concentrations tested whereas the percentage of floating conidia was close to 100% for all concentrations.

FIG 1.

The percentage of floating conidia inside the dialysis membrane (DM) tube normalized to the initial inoculum of the wild-type (AZN8196) and azole-resistant (V52-35) A. fumigatus isolates with anidulafungin dosing regimens with different fC_max values and t_1/2 of 18 h.

FIG 2.

In vitro concentration-effect relationship of anidulafungin against aberrant mycelia of the wild-type (AZN8196) and azole-resistant (V52-35) A. fumigatus isolates assessed using the height covering the dialysis membrane tubes in the in vitro PK/PD model after 72 h.

Pharmacodynamic differences were observed between the two susceptible strains. The percentage of floating conidia was smaller at higher concentrations for the azole-susceptible AZN8196 isolate (<50% floating conidia at fC_max of ≤0.11 μg/ml) than the percentage for the azole-resistant V52-35 isolate (<50% floating conidia at fC_max of ≤0.22 μg/ml), indicating increased antifungal susceptibility of AZN8196 (Fig. 1). Similarly, the concentration-time curve for AZN8196 was on the left of the concentration-time curve for V52-35, with the aberrant mycelia covering half of the height of the DM tube (∼7 cm) at a lower anidulafungin concentration of AZN8196 than of V52-35 (0.10 versus 0.18 μg/ml, P < 0.001) (Fig. 2).

Dose fractionation studies.

When the simulated anidulafungin animal dose of 20 mg/kg every 24 h (q24) was split into two doses of 10 mg/kg q12 resulting in similar free-drug AUCs from 0 to 24 h (fAUC_0–24s) but different fC_maxs, the heights of aberrant mycelia on the DM tube were similar (10.8 cm versus 10.9 cm). In contrast, with the simulated dosing regimen of 10 mg/kg q24, which results in lower fAUC values but similar fC_maxs with 10 mg/kg q12, the height was smaller (7.8 cm versus 10.9 cm) (Fig. 3). When micafungin and caspofungin were tested in the same model, the same phenomenon and results were found (data not shown). Thus, the AUC_0–24 best describes the activity of echinocandins.

FIG 3.

The AUC_0–24/MEC ratio is the driving PK/PD index for anidulafungin. The regimen with 20 mg/kg once daily (left photo) exhibited an effect similar to that of a regimen of 10 mg/kg twice daily (middle photo) and an effect lower than the regimen of 10 mg/kg once daily (right photo). The 20-mg/kg q24 regimen had an fAUC value similar to but an fCmax higher than that of the 10-mg/kg q12 regimen; the 10-mg/kg q12 regimen had an fC_max similar to but an fAUC higher than that of the 10-mg/kg q24 regimen. Similar results were also found for micafungin and caspofungin.

In vitro and in vivo exposure-effect relationship.

The in vitro exposure-effect anidulafungin relationship followed a sigmoid curve (R² = 0.92; hillslope, 1.96) similar to that of the animal model (R² = 0.85; hillslope, 1.34) (F test, P = 0.29) (Fig. 4). Accordingly, the in vitro fAUC_0–24 (95% confidence interval [CI]) corresponding to a 50% maximal net effect (PK₅₀) was 2.21 (1.81 to 2.71) mg · h/liter, close to the in vivo PK₅₀ of 2.62 (1.88 to 3.65) mg · h/liter fAUC_0–24 (F test, P = 0.41). When each isolate was analyzed separately, the in vitro PK₅₀ for the wild-type isolate A. fumigatus AZN8196 was lower than the PK₅₀ for the azole-resistant A. fumigatus V52-35 in vitro, with 1.80 (1.33 to 2.42) mg · h/liter versus 2.83 (1.96 to 4.11) mg · h/liter (P = 0.044), and in vivo, with 1.88 (1.4 to 2.53) and 4.05 (2.03 to 7.11) (P = 0.01), respectively.

FIG 4.

In vitro-in vivo correlation of exposure-effect relationship of anidulafungin against A. fumigatus. In vivo effects of high anidulafungin exposures (>100 fAUC/MEC) could not be assessed because of toxicity problems whereas it was determined that in vitro effects reached a maximal level at 270 fAUC/MEC.

In vitro PK/PD relationship.

Despite having the same MECs, the two isolates behaved differently in vitro and in vivo with one 2-fold dilution difference in exposure required for the 50% maximal effect. The other antifungal susceptibility testing methods detected a difference, but this difference was two 2-fold dilutions. A one 2-fold dilution difference was found using a 70% cutoff for the XTT method (the XTT MICs were 0.002 and 0.004 μg/ml for AZN8196 and V52-35, respectively). The in vitro PK/PD relationship for each method followed a sigmoid curve, with an R² of 0.89 to 0.92 (Fig. 5). The EI₅₀s (95% CI) for CLSI, EUCAST, Etest, and XTT endpoints were 132 (103 to 169), 66 (51 to 85), 35 (22 to 56), and 410 (254 to 659), respectively.

FIG 5.

In vitro PK/PD relationship of anidulafungin against A. fumigatus using different methods for assessing in vitro activity of anidulafungin.

DISCUSSION

A new marker for the activity of echinocandins was found utilizing the property of Aspergillus mycelia to attach on the dialysis membrane after exposure to echinocandins in a dose-response manner. This marker was easily quantified, reproducible, and amenable for PK/PD studies. The results with the new marker were comparable to those obtained from an animal model, with the new marker correlating with in vivo survival.

Echinocandins are fungicidal against rapidly growing hyphal tips of Aspergillus spp. and fungistatic against subapical regions of hyphae, mainly because chitin synthesis compensates the damage by echinocandins (7). Scanning electron microscopy showed that the debris resulting from the tip lysis adhered into a ceramic porous aluminum oxide support (8). Fluorescent staining with Syto9-propidium iodide revealed the presence of nucleic acids in the debris. Cellulose is a biopolymer that is used in blotting and filtration experiments (7). DNAs can bind to cellulose surfaces nonspecifically under typical aqueous conditions (9). Thus, it is very likely that the mechanism of the attachment of the aberrant mycelia observed in the present study on the DM of the PK/PD model is via the interaction between the cellulose of the DM and the nucleic acids released from the lysed hyphal tips in the presence of the echinocandins. Since mycelia may contain equal amounts of both lysed and intact hyphal tips (∼50:50) (8), the new marker is sensitive in detecting mycelia with only half-lysed hyphal tips within a mycelium.

The attachment of these mycelia on the DM tube was apparent within 24 h, and their number was inversely related to the number of floating conidia inside the DM tube. This phenomenon was concentration dependent since the number of attached aberrant mycelia on the DM tube was higher, and the number of floating conidia was lower at higher concentrations of anidulafungin. A clear concentration-dependent effect of echinocandins was previously found on ceramic porous material after 14 h of incubation when the diameter of the microcolonies (which correspond to the aberrant mycelia attached on the DM tube in the present study) was around 100 μm (8), which can be detected by naked eye. In agreement with the present study in which no aberrant mycelia attached on DM tube were found for the resistant isolate, when microcolonies of a caspofungin-resistant isolate were stained with Syto9-propidium iodide, tip lysis was rarely observed at concentrations sufficient to cause lysis in susceptible strains (8). In the previous study >50% cell lysis was found at 0.06 to 0.25 μg/ml of anidulafungin, which correlates with the large heights of aberrant mycelia attached on the DM tube (>7 cm of the total 12.3 cm of the DM tube). No paradoxical effect was observed in the present study, possibly because such an effect is usually observed at anidulafungin concentrations between 1.3 and 19.1 μg/ml, which were not tested in the present study (10).

The second interesting observation was that some conidia of the initial inoculum did not germinate and form mycelia but remained floating inside the DM tube. This indicates a second effect of echinocandins resulting in inhibition of germination of some conidia of the initial inoculum. The remaining conidia will germinate, and echinocandins will exert the classical mode of action by inhibiting 1,3-b-d-glucan synthase, causing the lysis of (some) hyphal tips of formed mycelia without reducing further the number of germinated conidia over time (8). This is in agreement with our observation that the height of aberrant mycelia did not change over time, but taking readings was easier after 48 h due to the larger size of abnormal mycelia attached on the DM tube. Previous studies have also shown that only 86% (79 to 89% among susceptible isolates) of untreated microcolonies were germinated after 14 h of incubation at 0.125 μg/ml of anidulafungin (8). Thus, echinocandins exert a growth-inhibitory effect at the conidial level in addition to the growth-inhibitory effect of subapical hyphal regions and the fungicidal effect of the apical hyphal regions. The new marker is sensitive in detecting the fungicidal action, and it can capture both growth-inhibitory effects since these effects are complementary, as found in the present study.

An intriguing observation was that the aberrant mycelia are attached only on one side of the DM tube. Since photosensitizing in fungi and particularly in Aspergillus species has been described previously (11), we assessed whether light was responsible for this phenomenon. However, aberrant mycelia were attached to the DM tube irrespective of light direction (data not shown), indicating that light was not responsible for this phenomenon. Quorum sensing may also play a role in such a way that the first mycelia are randomly attached on any side of the DM tube and subsequently attract other mycelia with quorum-sensing molecules (12). Mycelia are attached vertically and not horizontally around the DM tube because pores at the edges of the DM are compressed during production (personal communication with C. Markee, Spectrum Laboratories, San Jose, CA), and no pores are available to adsorb the extracellular material released due to echinocandin action. However, this phenomenon needs further investigation.

In vitro echinocandin activity is assessed with the MEC, which is subjective and demanding since it requires microscopic observation of all wells and detection of the shift from healthy to aberrant mycelia, which is not always clear and unambiguous. Furthermore, the MEC is not quantitative since it assesses the presence or absence of aberrant mycelia only. As shown in the present study, there is a concentration-dependent activity of echinocandins at supra-MEC concentrations, with more aberrant mycelia formed at higher concentrations that could not be captured by the MEC. In fact, these concentration-dependent activities were different for the two susceptible strains even though they had the same CLSI and EUCAST MECs. The gradient concentration test could detect these differences probably because of the 1.5-fold dilution step and the easier quantification of growth on agar than in the wells of microtiter plates. The XTT method also captured these differences although the difference in metabolic activities between the MICs of the two strains was only 20%. However, both the gradient concentration and XTT methods detected a two 2-fold dilution difference in MICs between the two strains when the exposure-effect relationship in the in vitro PK/PD model using the new marker showed a one 2-fold dilution difference, as found in animals. Further studies using more strains with different in vitro susceptibilities tested in animals are required in order to explore whether the CLSI and the EUCAST methods capture in vivo significant pharmacodynamic differences compared to results with the other methods.

In a rabbit model of pulmonary aspergillosis, the PK/PD index associated with the anidulafungin effect using residual fungal burden, lung weight pulmonary infarct score, and survival could not be clearly distinguished (13), emphasizing the difficulty of in vivo PK/PD studies with echinocandins. In a nonneutropenic murine model of disseminated aspergillosis, the AUC_0–24/MEC tended to correlate better with survival (5). However, the C_max/MEC was the driving PK/PD index for caspofungin in a neutropenic murine model of invasive pulmonary aspergillosis using PCR to assess pulmonary fungal burden (4). The same PK/PD index was found for micafungin in a persistently neutropenic rabbit model of invasive pulmonary aspergillosis (14). Difficulties in reliably assessing in vivo effects of echinocandins together with the different models (route of infection, immunosuppression, and concomitant therapy) used may hinder the driving PK/PD index. The activity of the drugs within the same pharmacological class is usually described by the same PK/PD index. PD studies with all three echinocandins using in vivo candidiasis models have demonstrated that the 24-h fAUC/MIC ratio is a good indicator of an exposure-response relationship and is considered the PK/PD index predicting therapeutic efficacy for all three echinocandins (15). Our results clearly showed that the activity of all three echinocandins against Aspergillus is also described by the PK/PD index fAUC/MEC.

In conclusion, a novel, easily quantified, sensitive, and reproducible marker for assessing echinocandin activity against Aspergillus spp. was described in the present study. The new marker correlated with in vivo outcome and clearly showed that the driving PK/PD index for all three echinocandins was the fAUC/MEC. The marker could be used in in vitro PK/PD studies in order to explore further the pharmacodynamics of this distinct class of antifungal agents against Aspergillus spp. and possibly other filamentous fungi and to optimize dosing regimens against invasive aspergillosis and other mold infections.

MATERIALS AND METHODS

Isolates and antifungal susceptibility.

Two clinical A. fumigatus isolates previously used in an animal model (5), a wild-type voriconazole-susceptible (AZN8196), and a voriconazole-resistant (V52-35) isolate with identical anidulafungin EUCAST (16) and CLSI (17) MECs of 0.03 and 0.015 μg/ml, respectively, and an anidulafungin-resistant strain (AFU79728) with a C2086T point mutation in the FKS1 gene (18) and EUCAST and CLSI MECs of >16 μg/ml were studied. Antifungal susceptibility testing was also performed using gradient concentration strips (Etest; bioMérieux) according to the manufacturer's instructions and the XTT method. The MIC was determined as the intersecting concentration of the elliptical inhibition zone at 50% inhibition (inhibition of confluent lawn of growth). For the XTT method, 50 μl of XTT solution (400 mg/ml XTT and 6.25 μM menadione [ΜΕΝ]) was added to each well after 24 h of incubation, according to the EUCAST microdilution method (19). Absorbance at 450/630 nm was measured after 1 h of incubation without removing the contents of the wells. The percentage of growth was calculated for each well by dividing the result by the absorbance of the drug-free control after subtracting the corresponding background absorbances. The XTT MIC was determined as the lowest concentration corresponding to <50% metabolism (19).

In vitro model.

A previously optimized two-compartment PK/PD dialysis/diffusion closed model (20) was used with a 10-ml dialysis membrane (DM) tube with 20-kDa molecular-mass-cutoff pore sizes (Spectra/Por Float-A-Lyzer G2; Spectrum Laboratories, Inc., Breda, The Netherlands). The DM tube was inoculated with 10³ CFU/ml in medium (RPMI 1640 medium, morpholinepropanesulfonic acid [MOPS], and 100 μg/ml chloramphenicol) and placed inside a glass beaker with medium that was diluted with fresh medium using a peristaltic pump at a rate equivalent to the elimination rate of anidulafungin. The system was incubated on a magnetic heating stirrer at 37°C. All experiments were performed in triplicate.

In vitro pharmacokinetics.

In order to validate the model, previously published in vivo results of an animal model were used (5). Specifically, anidulafungin dosages of 2.5, 5, 10, 20, and 40 mg/kg once daily [OD] resulted in total maximum mouse plasma concentrations (C_max) of 3, 7.9, 10.7, 22.2, and 49.5 μg/ml and AUC_0–24 values of 46.5, 93, 141.4, 326.3, and 802.7 mg · h/liter, respectively. An average half-life of 18 h was simulated in the in vitro PK/PD model targeting fC_max values, taking into account protein binding of 99% (15), i.e., 0.03, 0.08, 0.11, 0.22, and 0.5 μg/ml. Pure anidulafungin (Pfizer, Inc., Groton, CT, USA) solution prepared in dimethyl sulfoxide (DMSO; Chem-Lab NV, Zedelgem, Belgium) at 100× the final concentration was added at the corresponding fC_max values in the in vitro model once daily for 3 days. A drug-free control was also included. Drug levels were determined at regular time intervals using a microbiological agar diffusion assay, as described previously (3). In addition, the fAUC_0–24 was also calculated for each dosing regimen.

In vitro pharmacodynamics.

Microscopic and macroscopic evaluation of the contents of the DM tube with anidulafungin showed that there were two types of fungal forms inside the DM tube: floating conidia that could be measured by quantitative cultures and aberrant mycelia attached to the DM tube that could be quantified by the vertical height of the mycelia covering the DM tube (Fig. 6). Different areas of the DM tube were checked, and only abnormal mycelia were found. The percentage of floating conidia inside the DM tube was assessed microscopically to ensure that no hyphae were present and quantitatively with CFU counts at each time point divided by the CFU counts at 0 h. The percentage of aberrant mycelia attached on the DM tube was calculated based on the height reached by aberrant mycelia (single isolated mycelia on the DM tube were not taken into account) on the DM tube after 3 days, divided by the total height (12.3 cm) of the DM tube (Fig. 6). Notably, mycelia were attached only on one side of the DM tube. To verify whether aberrant mycelia attached to the DM tube were also observed for the other two echinocandins, a dosing regimen using an fC_max of 1 μg/ml of caspofungin and 0.11 μg/ml of micafungin with a 12-h half-life was tested in the in vitro PK/PD model. The coefficient of variation of this new marker was assessed in replicate experiments.

FIG 6.

Three examples of the degree of aberrant mycelia attached to the dialysis membrane (DM) tubes and floating conidia inside the DM tubes. The percentage of aberrant mycelia can be quantified using the height in the tube (height covered by aberrant mycelia/total height of the DM tube), and the percentage of floating conidia can be quantified with CFU counts (number of CFU after 72 h/number of CFU at 0 h). In the tube on the left, no aberrant mycelia are attached to the DM tube whereas 100% of the initial inoculum was detected by CFU counts and microscopic observation as floating inside the DM tube. In the middle tube, 65% of the total height of the DM tube was covered by aberrant mycelia whereas 41% of the initial inoculum was detected by CFU counts as floating inside the DM tube. In the tube on the right, 90% of the total height of DM tube was covered by aberrant mycelia whereas only 17% of the initial inoculum was detected by CFU counts as floating inside the DM tube. Heights of the attached mycelia are indicated on the figure. The fAUC_0–24s were calculated using the actual measured concentrations.

Dose fractionation studies.

To differentiate whether the AUC_0–24/MEC or C_max/MEC is the PK/PD index that best describes anidulafungin activity, three animal dosing regimens, namely, 20 mg/kg given once daily (A), 10 mg/kg given twice daily (B), and 10 mg/kg given once daily (C) with the same (A and B) or lower (C) AUC_0–24 values and the same (B and C) or higher (A) C_max values were tested against AZN8196. Similar experiments were performed for micafungin and caspofungin targeting simulated micafungin exposures with fC_maxs of 0.22 μg/ml q24 (A), 0.11 μg/ml q12 (B), and 0.11 μg/ml q24 (C) and caspofungin exposure with fC_maxs of 0.5 μg/ml q24 (A), 0.25 μg/ml q12 (B), and 0.25 μg/ml q24 (C). If the AUC_0–24/MEC is the important PK/PD index, one should expect an activity of A = B < C. If the C_max/MEC is the important PK/PD index one should expect an activity of A > B = C.

In vitro-in vivo correlation.

The in vitro relationship of percent aberrant mycelia-fAUC_0–24 after 3 days of treatment and the in vivo survival-fAUC_0–24 after 14 days of treatment with anidulafungin for experimental aspergillosis (5) were analyzed with nonlinear regression analysis using the E_max model described by the equation E = (E_max − E_min) × PKⁿ/(PKⁿ + PK_50ⁿ) + E_min, where E_max and E_min are the maximum and minimum effects, respectively, PK is the pharmacokinetic parameter fAUC_0–24, PK₅₀ is the PK corresponding to 50% of E_max − E_min, and n is the hillslope. The E_max and the E_min were shared among the data sets. The in vitro and in vivo PK₅₀ values and n were compared with the F test.

Furthermore, the in vitro PK/PD relationship of anidulafungin was explored by analyzing the in vitro relationship of percent aberrant mycelia-fAUC_0–24/MEC using the EUCAST and CLSI MECs and fAUC_0–24/MIC using the Etest and XTT MICs. The data were analyzed with the E_max model as described above, and the exposure index fAUC_0–24/MEC associated with 50% activity (EI₅₀) was calculated for each method.