Comparative pharmacodynamics and dose optimization of liposomal amphotericin B against Candida species in an in vitro pharmacokinetic/pharmacodynamic model

ABSTRACT

As comparative pharmacokinetic/pharmacodynamic (PK/PD) studies of liposomal amphotericin B (L-AMB) against Candida spp. are lacking, we explored L-AMB pharmacodynamics against different Candida species in an in vitro PK/PD dilution model. Eight Candida glabrata, Candida parapsilosis, and Candida krusei isolates (EUCAST/CLSI AMB MIC 0.125–1 mg/L) were studied in the in vitro PK/PD model simulating L-AMB C_max = 0.25–64 mg/L and t_1/2 = 9 h. The model was validated with one susceptible and one resistant Candida albicans isolate. The C_max/MIC-log₁₀CFU/mL reduction from the initial inoculum was analyzed with the E_max model, and Monte Carlo analysis was performed for the standard (3 mg/kg with C_max = 21.87 ± 12.47 mg/L) and higher (5 mg/kg with C_max = 83 ± 35.2 mg/L) L-AMB dose. A ≥1.5 log₁₀CFU/mL reduction was found at L-AMB C_max = 8 mg/L against C. albicans, C. parapsilosis, and C. krusei isolates (MIC 0.25–0.5 mg/L) whereas L-AMB C_max ≥ 32 mg/L was required for C. glabrata isolates. The in vitro PK/PD relationship followed a sigmoidal pattern (R² ≥ 0.85) with a mean C_max/MIC required for stasis of 2.1 for C. albicans (close to the in vivo stasis), 24/17 (EUCAST/CLSI) for C. glabrata, 8 for C. parapsilosis, and 10 for C. krusei. The probability of target attainment was ≥99% for C. albicans wild-type (WT) isolates with 3 mg/kg and for wild-type isolates of the other species with 5 mg/kg. L-AMB was four- to eightfold less active against the included non-C. albicans species than C. albicans. A standard 3-mg/kg dose is pharmacodynamically sufficient for C. albicans whereas our data suggest that 5 mg/kg may be recommendable for the included non-C. albicans species.

INTRODUCTION

Infections due to Candida spp. have increased significantly over the past few years and particularly during the COVID-19 pandemic (1). Candida albicans has been the leading cause of candidiasis, but in recent studies, >50% of Candida infections involves non-C. albicans species such as Candida parapsilosis, Candida glabrata, Candida krusei, and recently also Candida auris which usually exhibit resistance to one or several antifungal drugs (2). Amphotericin B (AMB) demonstrates uniform and potent activity against Candida spp. Although the epidemiological cutoff values of AMB against Candida spp. are usually 1–2 mg/L, time-kill studies showed marked species-specific differences with stronger and faster fungicidal activity found against C. albicans compared with other Candida species (3). Despite MICs below the susceptibility breakpoint of 1 mg/L, Candida lusitaniae is regarded intrinsically resistant to amphotericin B: (i) due to a higher spontaneous mutation rate and (ii) because amphotericin B is less fungicidal against this species compared with that against others (4). This raises the question whether species-specific differences of AMB fungicidal activity are clinically relevant also for the activity against other species. Although several preclinical pharmacokinetic/pharmacodynamic (PK/PD) studies of conventional AMB against Candida spp. exist demonstrating species-specific differences (5–7), there are only few pharmacodynamic studies for liposomal amphotericin B (L-AMB) that is mostly used for the treatment of Candida infections and mainly studying C. albicans isolates (8, 9). Thus, it is unknown whether the species-specific differences of AMB cidality against non-C. albicans species also exist for L-AMB, whether they are confirmed in dynamic models simulating in vivo pharmacokinetics, and whether they can be overcome by the higher drug exposures attained with L-AMB.

The approved L-AMB doses are 3 mg/kg for empiric therapy and 3–5 mg/kg for systemic infections including Candida infections (10). As there is no randomized clinical trial assessing the efficacy of the two doses, it is not clear which of the two later doses are more effective against which Candida infections. Recently, an in vitro PK/PD model indicated that 5 mg/kg of L-AMB is more effective than the standard dose of 3 mg/kg against C. auris which usually has higher MICs than other Candida spp. (11). Animal models showed too that higher L-AMB doses are required for C. krusei, C. glabrata, and C. lusitaniae compared with C. albicans experimental infections (9, 12, 13).

Therefore, in this study, we explored the pharmacodynamics of L-AMB against C. albicans, C. glabrata, C. parapsilosis, and C. krusei isolates, using an in vitro PK/PD dilution model simulating L-AMB pharmacokinetics, after validating it based on animal data of experimental candidiasis, and we subsequently determined the optimal doses for each species.

RESULTS

Comparison of in vitro PK/PD model with animal data

Obtained L-AMB C_maxs in the IC were within 20% of the target C_max of 0.125–128 mg/L, with half-lives of 7–11 h. The in vitro C_max/MIC-48 h log₁₀CFU/mL relationship followed a sigmoid curve (R² = 0.89). An MIC value of 32 mg/L was used for calculation of C_max/MIC ratios for the AMB-resistant C. albicans isolate SSI-2699. The mean C_max/MIC (95% CI) associated with stasis was 2.1, very close to that required for stasis in mouse kidneys (approximately 1.6–3.8 C_max/MIC). The latter was calculated based on L-AMB concentration at the site of infection, that is, the renal parenchyma. The kidney AUC associated with stasis was ~10 mg.h/L for an isolate with MIC 0.25 mg/L, i.e., 40 AUC/MIC. Given the 10.55–24.37 AUC/C_max ratio in mouse kidney, a C_max/MIC of 1.6–3.8 can be determined (8) (Fig. 1).

Fig 1.

In vitro PK/PD relationship of L-AMB for an AMB-susceptible (K1) and AMB-resistant (SSI-2699) C. albicans isolates tested in the in vitro PK/PD model using the 48-h change in log₁₀CFU/mL vs C_max/MIC compared with the initial inoculum.

Pharmacokinetics

The time-concentration profile of L-AMB is shown in Fig. 2. The pharmacokinetic parameters of L-AMB were well simulated in the in vitro model with an average half-life t_1/2 of 7.7 (5.3–13.9) h and with mean ± SD L-AMB C_max concentrations of 0.27 ± 0.07, 0.74 ± 0.09, 6.78 ± 0.73, 26.94 ± 1.58, and 49.45 ± 4.12 mg/L for target C_max 0.25, 1, 8, 32, and 64 mg/L, respectively.

Fig 2.

Representative time-concentration profile of simulated q24h L-AMB dosing regimens in the in vitro PK/PD model for Candida isolates. Data represent drug levels in the in vitro model (solid lines) and the respective target values (broken lines). Error bars represent standard errors.

Pharmacodynamics

Candida glabrata

After 48 h of incubation, drug-free controls grew from 4.18 ± 0.24 log₁₀CFU/mL to 6.94 ± 0.37 log₁₀CFU/mL. Following L-AMB exposure, a 2.5-log₁₀CFU/mL decrease of initial inoculum was observed at simulated exposures with C_max ≥ 8 mg/L for the isolate with EUCAST/CLSI AMB MIC 0.125/0.25 mg/L, while for the isolates with EUCAST/CLSI AMB MIC 0.5/1 and 1/1 mg/L, initial log₁₀CFU/mL decreased only for simulated exposures with C_max 32 and 64 mg/L, respectively (Fig. 3).

Fig 3.

Time-kill curves in the in vitro PK/PD model for each simulated L-AMB dosing regimen against the four C. glabrata isolates with increasing EUCAST/CLSI AMB MICs tested, using 10⁴ CFU/mL as initial inoculum. Horizontal dotted lines represent the limit of detection.

Candida parapsilosis

C. parapsilosis grew from a mean ± SD of 3.90 ± 0.21 log₁₀CFU/mL at t = 0 h to 7.10 ± 0.10 log₁₀CFU/mL at t = 48 h in drug-free controls. L-AMB produced 2.3 log₁₀CFU/mL reduction of initial inoculum against C. parapsilosis 1 with low EUCAST/CLSI AMB/ MIC 0.25 mg/L at C_max 8 mg/L, while a 0.8 log₁₀CFU/mL reduction of initial inoculum was observed for C. parapsilosis 2 with slightly higher EUCAST/CLSI AMB MIC 0.5 mg/L at C_max ≥ 32 mg/L (Fig. 4).

Fig 4.

Time-kill curves in the in vitro PK/PD model for each simulated L-AMB dosing regimen against the two C. parapsilosis and the two C. krusei isolates with increasing EUCAST/CLSI AMB MICs tested, using 10⁴ CFU/mL as initial inoculum. Horizontal dotted lines represent the limit of detection.

Candida krusei

Εach drug-free control grew by >2.5 log₁₀CFU/mL, namely, from 3.93 ± 0.01 log₁₀CFU/mL to 7.40 ± 0.10 log₁₀CFU/mL. A ≥2.4 log₁₀CFU/mL reduction of initial inoculum was obtained at C_max 8 mg/L for C. krusei 1 with EUCAST/CLSI AMB MIC 0.5/0.5 mg/L, whereas no log₁₀CFU/mL reduction of initial inoculum was found for C. krusei 2 with EUCAST/CLSI AMB MIC and 1/1 mg/L (Fig. 4). When the actual log₁₀CFU were analyzed, time-kill curves were shifted upwards by ~1log₁₀CFU due to a maximal 20× increase of the volume of the IC at 48 h (Fig. S1). For those isolates that regrew after initial killing, the MICs were similar to initial MICs indicating no emergence of resistance.

PK/PD analysis

The in vitro exposure-effect relationship for Candida isolates followed a sigmoid curve (R² = 0.60–0.94). Mean EUCAST/CLSI C_max/MIC (95% CI) associated with stasis was 24(18–32)/17(10–25) for C. glabrata isolates, 8 (5–15) for C. parapsilosis, and 10 (5-18) for C. krusei isolates. Notably, the in vitro 48-h PK/PD targets of L-AMB were approximately four- to eightfold higher than those required for C. albicans isolates [mean 2.1 (0.5–3.9) C_max/MIC, 95% CI) (Fig. 5). Similar PKPD targets were found when log₁₀CFU were analyzed taking into account the dilution of CFUs (Fig. S1).

Fig 5.

In vitro PK/PD relationship of L-AMB against [A] C. glabrata, [B] C. parapsilosis, and [C] C. krusei isolates tested in the in vitro model using CLSI MICs. Horizontal dotted lines represent stasis (no change in log₁₀CFU/mL compared with initial inoculum).

Probability of PK/PD target attainment

The probability of target attainment for the in vitro static PK/PD target with the standard and higher dose of L-AMB for Candida spp. is shown in Fig. 6. The probability of target attainment (PTA) for stasis with intravenous (i.v.) dose of 3 mg/kg 24h was >99% for C. albicans isolates with EUCAST/CLSI AMB MICs ≤ 2 mg/L validating the in vitro model as L-AMB is commonly used effectively to treat infections by wild-type C. albicans isolates. For the other Candida species, the PTA was >99% with the 5 mg/kg q24h i.v. for wild-type isolates.

Fig 6.

Probability of target attainment (PTA) for 5,000 patients receiving either standard (3 mg/kg q24h i.v.) or higher (5 mg/kg q24h i.v.) L-AMB dose were simulated with Monte Carlo analysis for different EUCAST and CLSI AMB MICs. Horizontal line corresponds to 95% PTAs. The epidemiological cut-off values for EUCAST (ECOFF) and CLSI (ECV) are shown for each species.

DISCUSSION

In the present study, an in vitro dilution PK/PD model was used to explore the in vitro pharmacodynamics of L-AMB against C. glabrata, C. parapsilosis, and C. krusei isolates. The in vitro model was validated using the same C. albicans isolate previously used in an animal neutropenic in vivo model of disseminated candidiasis, with good correlation of both the pharmacokinetic and pharmacodynamic parameters obtained after 48 h of treatment. In particular, the in vitro stasis of L-AMB (2.1 C_max/MIC) was very close to its in vivo activity in mouse kidneys (2.2 C_max/MIC), taking into account L-AMB concentration at the site of infection, the renal parenchyma (8). The PK/PD target of 2.1 C_max/MIC resulted in high PTA (>99%) with the standard dose of 3 mg/kg for wild-type C. albicans isolates with CLSI/EUCAST AMB MIC ≤ 2/1 mg/L. This is line with proven clinical activity of L-AMB against C. albicans infections (90% success rate for C. albicans isolates) (14) further validating our in vitro model. With the caveat that only few isolates per species were tested due to similar AMB MICs that most isolates exhibited, L-AMB was four- to eightfold less active against the other Candida species compared with C. albicans. The dose of 5 mg/kg will be needed to cover the wild-type population of C. glabrata and C. krusei (CLSI ECV 2 mg/L). Even for wild-type C. parapsilosis isolates (CLSI ECV 1 mg/L), the 3-mg/kg dose will be just enough whereas the 5-mg/kg dose will be needed for a higher PTA > 99% similar to that for C. albicans wild-type isolates. Of note, EUCAST ECOFF for those species is 1 mg/L which can be marginally covered with the standard dose of 3 mg/kg only for C. parapsilosis. Considering the variation in the PK/PD target for C. parapsilosis, the 5-mg/kg dose will cover EUCAST wild-type isolates. Thus, the 5 mg/kg L-AMB covers sufficiently wild-type populations for all three non-C. albicans species. Similar results were obtained by our group for C. auris isolates (11) for which L-AMB was approximately fourfold less active compared with C. albicans isolates (9 versus 2.1 C_max/MIC for stasis), favoring the use of higher L-AMB dosing (5 mg/kg) for isolates with MIC up to 2 mg/L.

Consistent with other in vitro studies, L-AMB exhibited a dose-dependent activity against C. albicans isolates. Specifically, increasing doses (≥2× MIC) led to a greater reduction in the fungal load (>2.5 log₁₀CFU/mL reduction), with its cidal activity being fast, even after 3 h of drug exposure in mice (9). However, at lower L-AMB doses, regrowth was observed at 6 h after the start of treatment, in agreement with previous in vivo studies (8), perhaps due to shorter post-antifungal effect observed at lower AMB concentrations (≤1× MIC) as previously described (15). This was also true for C. parapsilosis and C. krusei isolates for which regrowth was also shown at L-AMB concentrations ≤ 2× MIC. No emergence of resistance was observed as the recovered isolates had the same MIC as the initial pre-exposed isolates. As for the other Candida species tested, most studies focus on the activity of the conventional deoxycholate AMB formulation, with AMB achieving smaller and slower killing against C. glabrata, C. parapsilosis, C. krusei, and C. tropicalis isolates, compared with C. albicans and C. dubliniensis [minimum fungicidal concentration (MFCs) 2 to >16 vs ≤1 mg/L and cidal activity reached at 4× MIC after 2 h vs >16 h, respectively] (3). There are no in vitro PK/PD data for L-AMB and non-C. albicans spp.. In vivo studies with L-AMB have shown a dose-dependent activity, with L-AMB daily doses required to reduced significantly fungal kidney burden (>1 log) compared with untreated control in neutropenic animal models being higher for C. krusei and C. glabrata isolates than for C. albicans isolates (8 and 7.5 vs 1 mg/kg, respectively) (9, 12, 13). In a comparative study, 100% survival was found with 2.5 mg/kg of L-AMB against C. albicans and 5 mg/kg against C. tropicalis experimental infection in neutropenic mice (16). Thus, although C. dubliniensis and C. tropicalis were not tested in the present study, based on previous in vitro and animal studies, it seems that the pharmacodynamics of those two species are similar to C. albicans and non-C. albicans species tested in the present study, respectively.

It should be emphasized that in most in vivo studies, animal models were neutropenic, as in the present in vitro model, since there were no neutrophils. Enhanced killing was previously found when L-AMB was combined with peritoneal macrophages although other studies showed significant reduction of neutrophil uptake of C. albicans (17). Such an effect may reduce differences among species and probably make current findings applicable only to neutropenic patients. Indeed, the efficacy of the standard dose of L-AMB in patients with persistent neutropenia in randomized double-blind clinical trial was lower than the efficacy in non-neutropenic patients (71.4% vs 90.3%, respectively) (14). Of note, in the latter trial, the efficacy of L-AMB against C. glabrata was numerically lower than that against C. krusei and lower than those against C. albicans and C. parapsilosis (80% vs 85.7% vs 86.7 vs 89.3%, respectively) and mycological persistence was higher for C. glabrata and C. krusei compared with C. albicans and C. parapsilosis (20% vs 10%–11%, respectively) although the number of patients infected by non-C. albicans spp. was too small to infer a firm conclusion. High-dose AMB therapy was associated with favorable response compared with low-dose AMB therapy of C. krusei infections in immunocompromised patients (18). The two L-AMB doses of 3 and 5 mg/kg were compared in a randomized controlled clinical trial of empirical therapy of febrile neutropenia, and they were not statistically significant different although the 5-mg/kg dose performed slightly better (42% vs 40% successful response, 2.5% vs 4.8% fungal infections, 29.6% vs 40% persistent fever, and 0% vs 1.2% death related to fungal infections) (19). As the PTA of the standard dose for non-C. albicans wild-type isolates is 10%–90%, some patients will respond to the standard dose. Thus, a clinical trial to prove superiority of the higher dose will require a large number of patients.

In conclusion, L-AMB was four- to eightfold less active against C. parapsilosis, C. glabrata, and C. krusei compared with C. albicans isolates. Our data suggest that a high dose of L-AMB (5 mg/kg q24h i.v.) may be needed to treat infections by wild-type C. parapsilosis, C. glabrata, and C. krusei isolates in neutropenic patients. Although the latter dose is within the doses recommended in international guidelines, these in vitro findings need to be verified in vivo.

MATERIALS AND METHODS

Isolates

Four C. glabrata isolates, two C. parapsilosis and two C. krusei isolates, with different EUCAST and CLSI AMB MIC ranging from 0.125 to 1 and 0.25 to 1 mg/L, respectively, were tested (Table 1). Isolates were chosen based on AMB MICs and randomly among isolates with the same AMB MICs in order to minimize potential bias in selecting similar isolates and n order to capture as much as pharmacodynamic variability. The in vitro model was validated using one C. albicans (kindly provided by Prof. David Andes, University of Wisconsin, USA), previously tested in an animal model of disseminated candidiasis (8) and one well-characterized AMB-resistant C. albicans isolate (20) with EUCAST/CLSI AMB MIC 0.25/0.25 and >16/>16 mg/L, respectively. Broth microdilution testing was performed in accordance with EUCAST E.Def 7.3 (21) and CLSI M27-A3 (22) in triplicate experiments.

TABLE 1.

In vitro susceptibility of Candida isolates tested in the present study

Isolates	Median (range) values in mg/L
	Amphotericin B (AMB)
	EUCAST	CLSI
C. albicans K1	0.25 (0.25)	0.25 (0.25)
C. albicans SSI-2699a	>16 (>16)	>16 (>16)
C. glabrata 1	0.125 (0.125–0.25)	0.25 (0.125–0.25)
C. glabrata 2	0.5 (0.25–0.5)	0.5 (0.25–0.5)
C. glabrata 3	0.5 (0.25–1)	1 (1)
C. glabrata 4	1 (0.5–1)	1 (0.5–2)
C. parapsilosis 1	0.25 (0.25)	0.25 (0.25)
C. parapsilosis 2	0.5 (0.5)	0.5 (0.5)
C. krusei 1	0.5 (0.5)	0.5 (0.5)
C. krusei 2	1 (1)	1 (1)

^a SSI, Statens Serum Institut.

The isolates were stored in normal sterile saline with 10% glycerol at −70°C. Twenty-four hours prior to the study, the organisms were revived by subculturing on Sabouraud dextrose agar plates supplemented with gentamicin and chloramphenicol (SGC2, bioMerieux) and adjusted after counting in a Neubauer hemacytometer to a final inoculum of 10⁴ CFU/mL, verified by quantitative cultures on SGC2 plates.

Antifungal drug and medium

Amphotericin B (AMB; Sigma-Aldrich, Athens, Greece) was supplied as pure powder for the broth microdilution testing with stock solutions of 5 mg/mL being prepared in sterile dimethyl sulfoxide (Carlo Erba Reactifs-SDS, Val de Reuil, France) and stored at −70°C until use. The lipid formulation of Amphotericin B was used (AmBisome, Gilead Science Inc.), reconstituted according to the manufacturer’s instructions to a final concentration of 4 mg/mL. RPMI 1640 medium with L-glutamine, without bicarbonate (Sigma-Aldrich, Athens, Greece) buffered to pH 7.0 with 0.165 M MOPS and supplemented with 100 mg/L chloramphenicol (AppliChem GmbH, Darmstadt, Germany), was used as the growth medium.

In vitro PK/PD model

A previously developed one-compartment PK/PD dilution model (23), consisting of a 250-mL culture vessel (conical glass flask) containing fresh RPMI-1640 medium to an initial volume of 5 mL, for each Candida isolate and L-AMB dosing regimen, was used to simulate human L-AMB pharmacokinetics and study its antifungal activity. As AMB is present mainly in the form of liposomes in human serum after L-AMB administration (24), total L-AMB serum concentrations were simulated in the in vitro model. The culture vessel is connected to a peristaltic pump (Minipuls Evolution, Gilson Inc.), adding fresh medium in order to dilute its content at a rate as the clearance of L-AMB in plasma. Empty liposomes without AMB provided by Gilead Science Inc. were also tested in order to exclude any non-AMB-specific activity of L-AMB. The internal compartment (IC) was covered with an aluminum foil to minimize light exposure and placed on a heated (37°C) magnetic stirrer, while its volume increased over time reaching ~100 mL at 48 h.

Pharmacokinetic analysis

L-AMB drug exposures with C_max 0.25–64 mg/L and an average half-life of 9 h were simulated in the in vitro model. Drug concentrations were added at the corresponding C_max values in the in vitro model once daily for 48 h. The model was not run longer as pharmacodynamics do not change after this time period (25). Drug levels were determined at 0, 3, 24, 27, and 48 h using a microbiological diffusion assay using a Paecilomyces variotii strain as previously described (26). Due to the non-linearity between inhibition zones and L-AMB concentrations > 1 mg/L, samples from the IC with expected values > 1 mg/L were first diluted so that the concentration would fall in the linearity range 0.03-1 mg/L of the bioassay. The lowest limit of detection was 0.03 mg/L for the partial growth inhibition zone (80%). Interexperimental variability was assessed in replicate experiments.

Pharmacodynamic analysis

Fungal load inside the IC of each L-AMB dosing regimen was evaluated at regular intervals with 500 µL sampling at 0, 3, 24, and 48 h followed by quantitative cultures on SGC2 plates. Plates were incubated at 30°C for 24 h, and colonies were counted at each dilution. Dilutions that yielded 10–50 colonies were used in order to determine the log₁₀CFU/mL at each timepoint. Time-kill curves were constructed by plotting log₁₀CFU/mL over time. In order to eliminate carry over phenomena, which are usually observed at ≥8–16× MIC, those dosing regimen samples were spread rather than spotted on the agar plates (27). For conditions where fungi grew after the initial killing, the MICs of regrown isolates were tested in order to check the emergence of resistance, in case MICs were increased >2 twofold dilutions than the MICs of the isolate.

Comparison of in vitro PK/PD model with animal data

The in vitro PK/PD model was validated using previously published in vivo results in a neutropenic model of disseminated candidiasis in mice infected with the same C. albicans K1 strain also used in the present study and treated intraperitoneally with increasing L-AMB doses 0.312–80 mg/kg once daily for 72 h (8). An AMB-resistant C. albicans isolate (SSI-2699) was also tested. L-AMB exposures with C_max 0.125–128 mg/L and an average half-life of 11 h were simulated in the in vitro model. The log₁₀CFU/mL and drug levels were determined at regular intervals. L-AMB exposure (change in log₁₀CFU/mL from t = 0 h vs C_max/MIC) after 48 h of incubation, associated with stasis (no log₁₀CFU/mL reduction compared to the initial inoculum), was calculated with the Emax model and compared with the in vivo PK/PD target associated with 1log kill effect after 3 days of treatment (1log₁₀ reduction compared to the initial inoculum), as described below. Two independent experiments were conducted.

PK/PD analysis

The PK/PD index C_max/MIC ratio was calculated for each simulated dose, isolate, and experiment. The drug exposure-response relationship, expressed with the 48 h log₁₀CFU/mL reduction compared with the start of therapy values vs C_max/MIC at each dosing regimen and isolate, was analyzed with non-linear regression analysis using the sigmoidal model with a variable slope (E_max model) described by the equation E = (E_max − E_min)*EIⁿ/(EI₅₀ + EIⁿ) +E_min where E is the log₁₀CFU/mL reduction (dependent variable), E_max and E_min is the maximum log₁₀CFU/mL, respectively, EI is the exposure index C_max/MIC, EI₅₀ is the exposure index C_max/MIC corresponding to 50% of E_max-E_min, and n is the slope of the dose-effect relationship (Hill coefficient). Because the volume of the in vitro PK/PD model increased over time, a similar analysis was performed using the actual log₁₀CFU after multiplying the CFU/mL with the volume at each timepoint. All data were analyzed using the statistics software package GraphPad Prism, version 5.0, for Windows (GraphPad Software, San Diego, CA). All experiments were repeated twice.

Monte Carlo simulations and analysis

In order to bridge the in vitro data with clinical outcome, Monte Carlo simulation analysis was performed for 5,000 patients receiving the standard (3 mg/kg q24 i.v.) as well as the higher 5 mg/kg q24 i.v. dose, achieving blood levels corresponding to a mean ± SD L-AMB C_max 21.87 ± 12.47 mg/L and 83 ± 35.2 mg/L (28), respectively. The percentage of patients attaining the in vitro C_max/MIC corresponding to stasis of L-AMB compared with the initial inoculum was calculated for Candida isolates with CLSI MICs 0.25–8 mg/L. Recently published MIC distribution data for C. albicans and the other Candida spp. with EUCAST (29) and CLSI (30) were used.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aac.00225-24.

Figure S1. aac.00225-24-s0001.docx.

E_max model for Candida spp.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.