In Vitro Interaction of Voriconazole and Anidulafungin against Triazole-Resistant Aspergillus fumigatus

Abstract

Voriconazole is the recommended drug of first choice to treat infections caused by Aspergillus fumigatus. The efficacy of voriconazole might be hampered by the emergence of azole resistance. However, the combination of voriconazole with anidulafungin could improve therapeutic outcomes in azole-resistant invasive aspergillosis (IA). The in vitro interaction between voriconazole and anidulafungin was determined against voriconazole-susceptible and voriconazole-resistant (substitutions in the cyp51A gene, including single point [M220I and G54W] and tandem repeat [34-bp tandem repeat in the promoter region of the cyp51A gene in combination with substitutions at codon L98 and 46-bp tandem repeat in the promoter region of the cyp51A gene in combination with mutation at codons Y121 and T289] mutations) clinical A. fumigatus isolates using a checkerboard microdilution method with spectrophotometric analysis and a viability-based XTT {2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino)carbonyl]-2H-tetrazolium hydroxide} assay within 2 h of exposure after 24 and 48 h of incubation at 35°C to 37°C. Fractional inhibitory concentration (FIC) indexes (FICis) were determined using different MIC endpoints and Bliss independence analysis performed based on the response surface calculation of the no-drug interaction. Significant synergistic interactions obtained based on measuring the FIC index were dependent on the MIC endpoint, in which FICs were inversely related to voriconazole and anidulafungin MICs and were influenced by the CYP51A genotype. A statistically significant difference was observed between FIC indexes of isolates harboring tandem repeat mutations and wild-type controls (P = 0.006 by one-way analysis of variance [ANOVA]), indicating that synergy is decreased in azole-resistant strains. Our results indicated that a combination of voriconazole and anidulafungin might be effective against infections caused by both azole-susceptible and azole-resistant A. fumigatus isolates, but the combination could possibly be less effective in voriconazole-resistant strains with high MICs. Studies in vivo and in vitro-in vivo correlation investigations are required to validate the potential synergy of voriconazole and anidulafungin.

INTRODUCTION

Voriconazole (VCZ) is an extended-spectrum triazole which affects the integrity of the fungal cell membrane by inhibiting ergosterol biosynthesis. Voriconazole is the recommended first-choice therapy for infections caused by Aspergillus species (1, 2). However, acquired resistance to azoles was recently described for Aspergillus fumigatus, which may hamper the efficacy of voriconazole (3).

To date, a wide range of mutations in A. fumigatus have been described to confer azole resistance (3), which commonly involves changes in the cyp51A gene, the target for azole antifungals (4, 5). The emergence of azole resistance has been documented with increasing reports of azole-resistant clinical A. fumigatus isolates in multiple European countries, Asia, and the United States (5–11). There is increasing evidence that azole resistance is associated with treatment failure (4, 11, 12), and in a recent Dutch survey, azole-resistant invasive aspergillosis (IA) carried a mortality rate of 88% (11). These clinical observations are supported by preclinical studies in animal models of IA (5, 11, 13–19), where the MIC was shown to have a major impact on the efficacy of voriconazole and posaconazole (15, 20). Evidence is accumulating that azole resistance may develop in our environment with the consequence that in up to two-thirds of patients with azole-resistant Aspergillus disease, there was no history of previous azole exposure (11). Therefore, there is an urgent need for new approaches to manage azole-resistant Aspergillus diseases.

Although combination therapy is presently not recommended for the primary therapy of IA, it may be an effective alternative approach for treatment of patients with azole-resistant Aspergillus disease (21, 22). Several studies have shown the potential of combining an echinocandin with voriconazole to improve outcomes in IA (23–33), but in a recent prospective randomized study, the combination of voriconazole and anidulafungin (AFG) was found not to be more effective than voriconazole monotherapy (34).

Anidulafungin is a cyclic lipopeptide antifungal agent of the echinocandins with in vitro and in vivo activity against Aspergillus spp. (35), which acts via inhibition of 1,3-β-d-glucan synthesis present only in fungal cell walls (36). However, the drug is not clinically licensed for the treatment of IA. Despite the failure to show a benefit of voriconazole and anidulafungin therapy in IA, this combination might be an option for patients with azole-resistant IA disease.

In this study, we investigated the in vitro antifungal activity of voriconazole either alone or in combination with anidulafungin against a collection of 25 clinical A. fumigatus isolates, including voriconazole-resistant isolates with various substitutions in the cyp51A gene and voriconazole-susceptible isolates, to determine the interaction between these two agents.

(Parts of these results were presented at 51st ICAAC [Interscience Conference on Antimicrobial Agents and Chemotherapy], Chicago, IL, 17 to 20 September 2011.)

MATERIALS AND METHODS

Fungal isolates.

A collection of 25 clinical A. fumigatus isolates was used in this study. Clinical isolates harbored various substitutions in the cyp51A gene, including isolates with single point (M220I and G54W) and tandem repeat (TR₃₄/L98H and TR₄₆/Y121F/T289A) mutations, and voriconazole-susceptible clinical isolates without mutations in the cyp51A gene were used as wild-type controls (Table 1.). All isolates were obtained from the fungus culture collection of the Department of Medical Microbiology, Radboud University Nijmegen Medical Centre, Nijmegen, the Netherlands. The cyp51A gene substitutions and morphological strain identification were confirmed by sequence-based analysis, as described previously (5). The isolates had been stored in 10% glycerol broth at −80°C and were revived by subculturing on Sabouraud dextrose agar (SDA) supplemented with 0.02% chloramphenicol for 5 to 7 days at 35°C to 37°C. All isolates were subcultured again on SDA for 5 to 7 days at 35°C to 37°C before preparation of the inoculum. Candida parapsilosis (ATCC 22019) and Candida krusei (ATCC 6258) were used for quality control in all experiments.

Table 1.

FIC indices based on 10% and 25% growth endpoints and Bliss independence results for VRC-susceptible and VRC-resistant A. fumigatus isolates^a

Endpoint and Aspergillus fumigatus isolate (CYP51A substitution)	No. of isolates	Geometric mean value (range)
		MICVRC (mg/liter)	MICAFG (mg/liter)	FIC index	CVRC (mg/liter)	CAFG (mg/liter)	Bliss independence response surface analysis
							Sum ΔEb	Mean ΔE	SEM
10% growth
No mutation	7	0.58 (0.25–2)	≥1 (≥1–≥1)	0.46 (0.16–1.01)	0.25 (0.13–0.50)	0.01 (0.01–0.25)	193.3 (30.39–352.6)	2.30 (0.36–4.2)	0.66 (0.39–1.1)
Single point mutation (M220I and G54W)	5	1.49 (0.13–4)	≥1 (≥1–≥1)	0.38 (0.05–1.03)	0.48 (0.13–2.00)	0.04 (0.02–0.06)	90.16 (−207.6–409.6)	1.07 (−2.47–4.88)	0.78 (0.41–1.2)
Tandem repeat mutation (TR34/L98H and TR46/Y121F/T289A)	13	8.94 (2–32)	≥1 (≥1–≥1)	0.33 (0.04–1.01)	2.18 (0.02–8.00)	0.04 (0.01–0.25)	95.74 (−335.5–692.8)	1.14 (−3.99–8.25)	0.53 (0.27–0.72)
25% growth
No mutation	7	0.40(0.25–1)	0.68 (0.03–≥1)	0.08 (0.02–0.54)	0.01 (0.01–0.13)	0.02 (0.02–0.06)	193.3 (30.39–352.6)	2.30 (0.36–4.2)	0.66 (0.39–1.1)
Single point mutation (M220I and G54W)	5	1.16(0.13–4)	0.61 (0.06–≥1)	0.10 (0.02–0.56)	0.02 (0.01–0.25)	0.04 (0.02–0.13)	90.16 (−207.6–409.6)	1.07 (−2.47–4.88)	0.78 (0.41–1.2)
Tandem repeat mutation (TR34/L98H and TR46/Y121F/T289A)	13	10.22(2–32)	0.09(0.02–≥1)	0.19 (0.01–1.03)	0.04 (0.01–4.00)	0.01 (0.01–0.13)	95.74 (−335.5–692.8)	1.14 (−3.99–8.25)	0.53 (0.27–0.72)

^aThe MICs of voriconazole and anidulafungin alone (MIC_VRC and MIC_AFG, respectively) and the concentrations of voriconazole (C_VRC) and anidulafungin (C_AFG) in combination are presented. Data are presented as geometric means and ranges.

^bBliss interaction.

Preparation of inoculum.

Conidial suspensions were harvested after isolates were subcultured on SDA at 35°C to 37°C 2× 5 to 7 days and were suspended in normal saline containing 0.025% Tween 20. Aspergillus inocula were then prepared spectrophotometrically and further diluted in normal saline in order to obtain a final inoculum concentration of 2 × 10⁵ to 5 × 10⁵ CFU/ml (37).

Antifungal agents.

Voriconazole and anidulafungin (Pfizer, Capelle aan den Ijssel, the Netherlands) were obtained as standard pure powders, and serial dilutions were prepared according to European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines (37).

Susceptibility and drug interaction testing.

Antifungal susceptibility (MICs and minimum effective concentrations [MECs]) and drug interaction testing were performed by using the EUCAST broth microdilution checkerboard (two dimensional, 8 by 12) method (37), utilizing XTT dye {2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino)carbonyl]-2H-tetrazolium hydroxide} (38–40). XTT (Sigma-Aldrich, St. Louis, MO) was dissolved in normal saline at concentrations of 0.5 mg/ml. Menadione (Sigma-Aldrich, St. Louis, MO) was initially dissolved in absolute ethanol at a concentration of 10 mg/ml and subsequently added to the above-mentioned XTT solutions at concentrations of 6.25 μM for each solution. The final concentrations of the antifungal agents ranged from 0.016 to 16 mg/liter for voriconazole and 0.008 to 0.5 mg/liter for anidulafungin. Aliquots of 50 μl of each drug at a concentration four times the targeted final concentration were dispensed into the wells of flat-bottom 96-well microtiter plates (Costar; Corning, NY). Trays were maintained for a period of less than 1 month at −70°C until the day of testing. After the microtitration trays were defrosted, 100 μl of the inoculum was added to each well, corresponding to a final concentration of 2 × 10⁵ to 5 × 10⁵ CFU/ml from each isolate. The microtiter plates were incubated at 35°C to 37°C for 48 h. Subsequently, 50 μl of the above-mentioned XTT-menadione solutions was added to each well, as previously described (40, 41). The microtitration plates were further incubated at 35°C to 37°C for 2 h in order to allow conversion of XTT to its formazan derivative. XTT conversion was measured as optical density (OD) with a microtitration plate spectrophotometric reader (Anthos htIII; Anthos Labtec Instruments, Salzburg, Austria) at 450 nm/630 nm. For each well, XTT conversion was calculated after subtraction of the background OD, which was the OD of a simultaneously incubated well with 200 μl of medium and 50 μl of XTT-menadione solution but no inoculum. Percentages of fungal growth were calculated for each well by dividing the XTT conversion of each well by the XTT conversion of the drug-free growth control well. All experiments were performed in three independent replicates, and the breakpoints reported previously by Verweij et al. were used for classifying voriconazole-susceptible and voriconazole-resistant isolates (3).

MIC and MEC determination.

The MIC of voriconazole was defined as the lowest concentration that completely inhibited growth compared with that of the drug-free well, as assessed by visual inspection. The MEC of anidulafungin was defined as the lowest concentration in which abnormal, short, and branched hyphal clusters were observed, in contrast to the long, unbranched hyphal elements that were seen in the growth control well (37). Because the voriconazole MIC corresponds to the lowest drug concentration corresponding to <10% growth and the MEC corresponds to the lowest concentration corresponding to <50% growth with the XTT assays, for the voriconazole-anidulafungin combination, both 10% and 50% growth endpoints in addition to the 25% growth endpoint were considered MIC endpoints.

Definitions for drug interaction modeling.

In order to assess the nature of in vitro interactions between voriconazole and anidulafungin, the data obtained as described above were analyzed using two different models. These models were nonparametric approaches of the following two no (zero)-interaction theories: the Loewe additivity (LA) and the Bliss independence (BI) theories (42–45).

The fractional inhibitory concentration (FIC) index (FICi) is defined as ∑FIC = FICA + FICB = C_A^comb/MIC_A^alone + C_B^comb/MIC_B^alone, where MIC_A^alone and MIC_B^alone are the MICs of drugs A and B when acting alone and C_A^comb and C_B^comb are the concentrations of drugs A and B at the isoeffective combinations, respectively (42). To determine synergistic and antagonistic interactions among all ∑FICs calculated for each isolate and replicate, the FIC index was determined as the ∑FIC_min (the lowest ∑FIC) or the ∑FIC_max (the highest ∑FIC) (42). The 10%, 25%, and 50% endpoints of fungal growth were used to assess pharmacodynamic interactions at different concentrations. In order to determine the nature of the interaction between voriconazole and anidulafungin, previously described cutoff values were used (46), in which an interaction was defined as synergistic if the FIC index was ≤1, additive if the FIC index was >1 to ≤1.25, and antagonistic if the FIC index was >1.25. These cutoff values were derived from experiments that investigated the voriconazole-echinocandin interaction (46). Furthermore, we compared our analysis with the commonly used FICi range of 0.5 to 4 that is generally recommended to define drug-drug interactions in combination studies of antifungal agents (21, 47, 48).

The BI was described by the equation I_ind = I_A + I_B − I_A × I_B, where I_ind is the predicted percentage of inhibition of an noninteractive theoretical combination, calculated based on the experimental percentages of inhibition (I_A and I_B) of each drug acting alone, respectively (43). In the three-dimensional plots, peaks above and below the zero plane indicate synergistic and antagonistic combinations, respectively, whereas the zero plane itself indicates no statistically significant interactions. The average sum of the three replicates of all Bliss interactions was used as a measure of the pharmacodynamic interactions for each strain.

Data analysis.

All data analyses were performed by using the software package GraphPad Prism, version 5.0, for Windows (GraphPad Software, San Diego, CA). The FICs among the different genotype groups were compared by analysis of variance (ANOVA) followed by a posttest for linear trends. The correlation between the mean FIC indexes and voriconazole and anidulafungin MIC endpoints was determined by Spearman's correlation coefficient (r); a P value of 0.05 was considered significant (two tailed).

RESULTS

The MIC and MEC characteristics of the 25 clinical A. fumigatus isolates used for the current study are shown in Table 1. The mean MICs of voriconazole (and ranges) based on 10% and 25% growth endpoints were 0.58 (0.25 to 2) mg/liter and 0.40 (0.25 to 1) mg/liter, respectively, for the voriconazole-susceptible (VCZ-S) isolates, whereas higher MICs were observed for isolates harboring single point mutations, 1.49 (0.13 to 4) mg/liter and 1.16 mg/liter (0.13 to 4), respectively, and tandem repeat mutations, 8.94 (2 to 32) mg/liter and 10.22 (2 to 32) mg/liter, respectively. Anidulafungin MIC endpoints based on the 50% growth endpoint were off scale for most of the isolates, and therefore, this growth endpoint was excluded from the analysis.

The mean values of FIC indexes based on 10% and 25% growth endpoints as well as BI response surface analysis results for different groups of A. fumigatus isolates with regard to substitutions in the cyp51A gene are also shown in Table 1, whereas Fig. 1 shows the distribution of FICs at each growth endpoint. None of the data sets analyzed had ∑FIC_maxs higher than 1.25, indicating that antagonism was not observed. Therefore, the FIC index corresponded to the ∑FIC_min. The lowest FIC index values found for isolates without a mutation in the cyp51A gene ranged between 0.16 and 1.01 based on the 10% growth endpoint and between 0.02 and 0.54 with the 25% growth endpoint, followed by isolates harboring single point and tandem repeat mutations, respectively. For isolates with the tandem repeat resistance mechanism (TR₃₄/L98H and TR₄₆/Y121F/T289A), the FIC index values averaged 0.33 (range, 0.04 to 1.01) based on the 10% growth endpoint and 0.19 (range, 0.01 to 1.03) with the 25% growth endpoint.

Fig 1.

Graphical distribution of mean and standard error of the mean of FIC indexes determined at 10% and 25% growth endpoints for 25 A. fumigatus isolates. None of the data sets analyzed had ∑FIC_maxs higher than 1.25, indicating that antagonism was not observed.

When analyzing interactions, considering the 10% and 25% growth endpoints, significant synergy (P ≤ 0.05) was found for all isolates, with mean FICi_mins of 0.42 and 0.12, respectively (Fig. 1). However, the wide distribution observed for mean FIC values of each growth endpoint indicated that for some strains, there appeared to be no synergism.

As shown in Fig. 2, the mean FICi based on the 25% growth endpoint did not differ significantly among voriconazole-susceptible isolates and those with single point mutations (M220I and G54W). However, a statistically significant difference was observed between isolates harboring tandem repeat mutations (TR₃₄/L98H and TR₄₆/Y121F/T289A) and wild-type controls (P = 0.006 by one-way analysis of variance [ANOVA]). Therefore, the dependence of the FIC index on resistance mechanisms indicates that synergistic interactions may be lost for the isolates with higher MICs of voriconazole. The consequence of this observation is that in isolates where voriconazole has no in vitro activity (MIC ≥ 8 mg/liter), the efficacy of the combination relies solely on anidulafungin.

Fig 2.

Mean and standard errors of the mean of FICs with respect to cyp51 substitutions for 25 A. fumigatus isolates, indicating that the FIC indexes are dependent on the type of mutation. The vertical bars indicate that the mean FIC indexes did not differ significantly among VCZ-S isolates and those with M220I and G54W mutations; a statistically significant difference was observed between isolates harboring TR₃₄/L98H and TR₄₆/Y121F/T289A mutations and wild-type controls (P < 0.05).

Furthermore, the results of FICi analysis are supported by response surface analysis using the BI no-interaction model for all isolates where the synergistic interactions in wild-type isolates were higher than those in the other two groups harboring CYP51A gene mutations for which some antagonistic interactions were observed. Bliss antagonism reflects additive/indifferent interactions by Loewe additivity. Thus, the presence of antagonistic interactions correlates with the reduction of Loewe synergistic interactions at the 25% growth endpoint. The selected interaction surface plots indicating synergy and antagonism for a voriconazole-susceptible A. fumigatus isolate (MIC of voriconazole, 0.25 mg/liter; MEC of anidulafungin, 0.03 mg/liter) and a voriconazole-resistant A. fumigatus isolate (MIC of voriconazole, 4 mg/liter; MEC of anidulafungin, 0.03 mg/liter) are shown in Fig. 3.

Fig 3.

Interaction surfaces obtained from response surface analysis of the Bliss independence no-interaction model for the in vitro combination of VCZ plus AFG against a VCZ-susceptible A. fumigatus isolate (MIC of VCZ, 0.25 mg/liter; MEC of AFG, 0.03 mg/liter) and a VCZ-resistant A. fumigatus isolate (MIC of VCZ, 4 mg/liter; MEC of AFG, 0.03 mg/liter). The x and y axes represent the efficacies of VCZ and AFG, respectively. The z axis is the percent ΔE. The zero plane represents Bliss-independent interactions, whereas the volumes above the zero plane represent statistically significantly synergistic (positive ΔE) interactions. The magnitude of interactions is directly related to ΔE. The different tones in three-dimensional plots represent different percentile bands of synergy. (A) Synergistic interaction. The mean ΔE ± standard error of the mean and sum ΔE were 3.23% ± 1.09% and 271%, respectively, after 48 h. (B) Antagonistic interaction. The mean ΔE ± standard error of the mean and sum ΔE were −2.47% ± 0.40% and −208%, respectively, after 48 h.

In comparison, Fig. 4 shows the interpretation of FIC indices, using two different cutoff values, in which the commonly used FIC index range of 0.5 to 4 indicated synergism for 38.1% and indifference for 61.9% of isolates, while the use of recently reported cutoff values (46) indicated synergism for 75.0% and additivity for 25.0% of isolates. Antagonism was not observed with either definition of the interaction.

Fig 4.

Interpretation of voriconazole and anidulafungin interactions for 25 A. fumigatus isolates utilizing two different definitions: the cutoff values proposed previously by Meletiadis et al. (46) (synergistic if the FIC index was ≤1, additive if the FIC index was >1 to ≤1.25, and antagonistic if the FIC index was >1.25) and the commonly used FIC index range of 0.5 to 4 proposed previously by Greco et al. (47) and which is generally recommended to define drug interactions in most combination studies of antifungal agents (synergistic if the FIC index was ≤0.5, indifferent if the FIC index was >0.5 to ≤4, and antagonistic if the FIC index was >4).

DISCUSSION

A number of studies have reported data on the efficacy of combination therapy against A. fumigatus. Most studies investigating combinations of azoles and echinocandins have shown a synergistic or additive interaction against Aspergillus spp. (24, 27, 29, 30, 33). Antagonism was not reported. The combination of voriconazole and an echinocandin in advanced invasive pulmonary aspergillosis in transiently neutropenic rats improved the therapeutic outcome (49). Notably, synergy was documented by the majority of studies when susceptibility testing endpoints were defined as a substantial inhibition of growth. For example, in a previous study by Shalit et al., caspofungin and itraconazole were studied alone and in combination against 31 clinical Aspergillus isolates (33). MICs and MECs were recorded, and synergy was calculated by using both endpoints. Synergy or synergy to additivity was found for 30 of 31 isolates by using MIC endpoints. With MEC endpoints, no synergy was found, and indifference was detected for 26 of 31 strains. In a previous study by Philip et al., significant synergy was noticed with regard to combinations of voriconazole and anidulafungin for 18/26 isolates, depending on the drug concentration and interaction definitions (32). Voriconazole in combination with anidulafungin has been shown to be efficient in treating infections caused by A. fumigatus in an immunosuppressed guinea pig model of IA (25, 49). We recently also found a synergistic interaction between voriconazole and anidulafungin in a model of disseminated IA when mice were infected with a voriconazole-susceptible isolate (50).

Although retrospective clinical studies indicated a benefit of combining an echinocandin, i.e., caspofungin, with voriconazole (28), a recent randomized prospective trial of voriconazole and anidulafungin showed no superiority to voriconazole monotherapy (34). This apparent discrepancy between this prospective clinical trial and retrospective trials and preclinical research may be due to methodological issues related to the prospective clinical trial (27). However, preclinical studies involved only wild-type isolates, and it can be assumed that the vast majority of patients enrolled in clinical studies would have suffered from invasive aspergillosis due to wild-type isolates.

In azole-resistant disease, combination therapy has potential benefit, as the reduced efficacy of the azole might be overcome by the concomitant administration of an echinocandin. In our murine model, we found that the interaction between voriconazole and anidulafungin was indifferent in mice infected with an A. fumigatus isolate with a voriconazole MIC of 4 mg/liter (50), which indicated that the drug interaction varied according to the susceptibility of the isolate to voriconazole. As only one azole-resistant isolate was investigated in the animal model, we used an in vitro interaction model to investigate this relationship in more detail using a larger collection of isolates and a wide range of voriconazole MICs. Furthermore, fitting an interaction model to the whole response surface and estimation have the additional advantage that confidence intervals of the interaction are obtained (44).

We found that synergistic drug interactions obtained for the FIC indexes were dependent on the MIC endpoints. Significant variations were observed in the FIC distributions using MIC endpoints. However, for some strains, there appeared to be no synergism (FIC > 1), which was dependent on the MIC of voriconazole. This variation in FIC index results could be explained largely by the CYP51A gene mutation and the associated voriconazole phenotype of the strain. In addition to the analysis with the nonparametric fractional inhibitory concentration model (FIC index), similar results were found when the data were analyzed using the response surface approaches of the Bliss independence (BI) no-interaction theory.

The statistically significant difference between isolates harboring tandem repeat mutations and wild-type controls (P = 0.006 by ANOVA) is in keeping with the observation in our in vivo model (50). FICs were inversely related to voriconazole and anidulafungin MICs and influences by CYP51A genotype.

The interpretation of data from in vitro interaction studies depends on the definition used for FIC calculation (21, 22, 46–48, 51–55), which can vary depending on the cutoff values used (Fig. 4). In our study, we used cutoff values to indicate that the interactions were synergistic if the FIC index was ≤1, additive if the FIC index was >1 to ≤1.25, and antagonistic if the FIC index was >1.25 (46), since an additivity range of 0.5 to 2 is more symmetrical than a range of 0.5 to 4. Furthermore, the cutoffs of 1 and 1.25 were previously investigated for drug interactions of voriconazole and anidulafungin against A. fumigatus and validated by an in vivo model (46). Interpreting our data by this definition indicated synergism for 75% of isolates and additivity for 25% of isolates. In comparison, the application of the generally used FIC index range of 0.5 to 4 (21, 47, 48) indicated synergism for 61.9% of isolates and indifference for 38.1% of isolates.

We used XTT for a more precise quantification of hyphal growth. It has been shown that the assessment of metabolic activity provides useful quantitative endpoints for in vitro studies of both azoles and echinocandins against Aspergillus spp. (38, 40, 41, 43).

The significant relationship between FICi and CYP51A genotype raises concern regarding if the combination of voriconazole and anidulafungin can be used in the management of azole-resistant disease. In the Netherlands, the TR₃₄/L98H mutation is highly prevalent (5, 11), and more recently, a TR₄₆/Y121F/T289A mutation was found in A. fumigatus isolates recovered from patients from multiple Dutch hospitals (56). This new resistance mechanism has characteristics similar to those of TR₃₄/L98H, indicating that it may also originate from the environment. These two resistance mechanisms correspond to the highest voriconazole MICs (8 and >16 mg/liter), and our results indicate that we can expect the least benefit from combination therapy with voriconazole and anidulafungin in patients infected by A. fumigatus strains harboring these resistance mechanisms. As the targets of azoles and echinocandins are unrelated, a lack of voriconazole activity may indicate that the efficacy of combination therapy relies solely on anidulafungin.

Evidence to support treatment choices for azole-resistant Aspergillus disease is scarce at present. Although the in vitro activity of echinocandins and amphotericin B appears unaffected in azole-resistant isolates, in vivo efficacy studies are lacking. Clearly, more research is warranted to explore treatment options in azole-resistant disease. Our results indicate that azole and echinocandin combination therapy should be used with great caution in patients with azole-resistant Aspergillus diseases.