ABSTRACT
Antifungal susceptibility testing of Aspergillus species has been standardized by both the Clinical and Laboratory Standards Institute (CLSI) and the European Committee on Antimicrobial Susceptibility Testing (EUCAST). Recent studies suggest the emergence of strains of Aspergillus fumigatus with acquired resistance to azoles. The mechanisms of resistance involve mutations in the cyp51A (sterol demethylase) gene, and patterns of azole cross-resistance have been linked to specific mutations. Studies using the EUCAST broth microdilution (BMD) method have defined wild-type (WT) MIC distributions, epidemiological cutoff values (ECVs), and cross-resistance among the azoles. We tested a collection of 637 clinical isolates of A. fumigatus for which itraconazole MICs were ≤2 μg/ml against posaconazole and voriconazole using the CLSI BMD method. An ECV of ≤1 μg/ml encompassed the WT population of A. fumigatus for itraconazole and voriconazole, whereas an ECV of ≤0.25 μg/ml was established for posaconazole. Our results demonstrate that the WT distribution and ECVs for A. fumigatus and the mold-active triazoles were the same when determined by the CLSI or the EUCAST BMD method. A collection of 43 isolates for which itraconazole MICs fell outside of the ECV were used to assess cross-resistance. Cross-resistance between itraconazole and posaconazole was seen for 53.5% of the isolates, whereas cross-resistance between itraconazole and voriconazole was apparent in only 7% of the isolates. The establishment of the WT MIC distribution and ECVs for the azoles and A. fumigatus will be useful in resistance surveillance and is an important step toward the development of clinical breakpoints.
Invasive aspergillosis (IA) is second only to candidiasis as the most important invasive mycosis affecting humans (5, 10, 18, 37, 40). Although several hundred species of Aspergillus have been described, relatively few are known to cause disease in humans. Aspergillus fumigatus remains the most common cause of IA, although the proportion of IA cases caused by this species has fallen from ∼90% of cases in the 1980s to between 50 and 60% of cases in the 1990s and into the 2000s (30, 36).
Antifungal susceptibility testing of Candida as an aid in guiding clinical treatment has gained wide acceptance (3, 8, 16, 20, 22, 24, 39). The development and application of in vitro susceptibility testing of Aspergillus spp. have lagged behind that for Candida due to the difficulty in determining a clinical role for this testing. Confounding the issue are numerous factors that play either a positive or negative role in determining the outcome of therapy (14, 26). Given the increasing number of antifungal agents with systemic activity against Aspergillus spp. (13, 15, 21, 38, 41, 42, 44, 48, 49), it is recognized that antifungal susceptibility testing of these opportunistic pathogens may be useful in guiding the selection of antifungal agents for the treatment of IA (1, 34, 42, 44, 47). This is especially true for itraconazole (11, 34) and the newer extended-spectrum triazoles posaconazole and voriconazole (42, 44).
In vitro antifungal susceptibility testing of azoles versus Aspergillus spp. has been standardized by both the CLSI (7) and European Committee on Antimicrobial Susceptibility Testing (EUCAST) (27). Both methods use a 96-well broth microdilution (BMD) format, RPMI 1640 broth medium, 48 h of incubation, and a MIC endpoint of complete inhibition of growth as determined by visual inspection. The EUCAST method employs a higher inoculum concentration (2 × 105 to 5 × 105 CFU/ml versus 0.4 × 104 to 5 × 104 CFU/ml) and additional glucose (2% versus 0.2% final concentration) in an effort to improve fungal growth. Using these methods, A. fumigatus isolates exhibiting high MICs to azole drugs have been described (2, 11, 19, 28, 29) and both methods have been shown to reliably detect those strains with defined azole resistance mechanisms (1, 4, 6, 17, 31-33, 42, 44).
Mechanisms of elevated azole MICs in A. fumigatus discovered thus far involve the cyp51A (lanosterol demethylase) gene and include mutations resulting in amino acid substitutions at glycine 54 (G54) (12, 29, 35, 44) at G448 (50), and at methionine 220 (M220) (31, 44) and an amino acid substitution of leucine for histidine at position 98 (L98H) together with a tandem repeat (TR) of a 34-bp sequence in the promoter of the cyp51A gene (33). Cross-resistance to itraconazole and posaconazole has been associated with the G54 substitution, whereas cross-resistance to voriconazole and ravuconazole has been associated with the G448 substitution (44, 50). Both the M220 substitution and the L98H TR produce a cross-resistance pattern to all four azoles (44).
Interpretive breakpoints based upon the correlation of in vitro data with clinical outcome have not been established for any Aspergillus-drug combination (14, 42, 44). In the absence of the necessary clinical data, one practical approach to the use of susceptibility testing data in detecting resistance or decreased susceptibility has been to define the wild-type (WT) distribution of MICs for the relevant drug-organism combinations (e.g., populations of organisms with no acquired resistance mechanisms) and set epidemiological cutoff values (ECVs) that would discriminate WT strains from those with acquired resistance mechanisms (25, 44-46). The clinical relevance of ECVs would still be uncertain, but ECVs could nonetheless serve as the foundation for the laboratory detection of acquired resistance (decreased susceptibility) and be used to monitor resistance development (45).
Rodriguez-Tudela et al. (44) employed the EUCAST BMD method to define the WT MIC distribution of four triazole antifungal agents (itraconazole, posaconazole, ravuconazole, and voriconazole) for A. fumigatus. They also used studies of resistance mechanisms in A. fumigatus to demonstrate that ECVs of ≤1 μg/ml for itraconazole, ravuconazole, and voriconazole and ≤0.25 μg/ml for posaconazole identified the WT strains and provided separation of the WT population from those strains with resistance mutations in the cyp51A gene (44). Definition of the WT distribution and ECVs for azoles and A. fumigatus is important in order to have a clear understanding of which MICs can be considered microbiologically susceptible (25, 44, 46). Furthermore, these studies showed that the MIC phenotype obtained by means of a standardized methodology was sufficient to identify nonsusceptible strains of A. fumigatus and their pattern of cross-resistance without the need for sequence analysis of cyp51A to detect the underlying mutation (44).
Similar studies using CLSI BMD to test itraconazole, posaconazole, ravuconazole, and voriconazole against 553 isolates of A. fumigatus confirmed the cross-resistance profiles reported by Rodriguez-Tudela et al. (44), indicating that both CLSI and EUCAST BMD methods may be used to sort out these different patterns (42). In the present study, we utilized a large, geographically diverse collection of A. fumigatus isolates tested against itraconazole, posaconazole, and voriconazole by the CLSI method to provide further documentation of the WT MIC distribution and ECVs for the azoles and A. fumigatus.
MATERIALS AND METHODS
Organisms.Between January 2005 and December 2007, 637 unique patient isolates of A. fumigatus were obtained from more than 60 different medical centers worldwide. All isolates were submitted to the Molecular Epidemiology and Fungus Testing Laboratory (MEFTL) at the University of Iowa College of Medicine as part of a global antifungal surveillance program. Each participating center was instructed to submit fungal isolates from sterile sites or from respiratory samples when they were determined to be significant pathogens by local criteria. The isolates were obtained from a variety of sources, including sputum, bronchoscopy, and tissue biopsy specimens. All isolates were identified by standard microscopic morphology and were stored as spore suspensions in sterile distilled water at room temperature until used in the study. Before testing, each isolate was subcultured at least twice on potato dextrose agar (Remel, Lenexa, KS) to ensure viability and purity. As a screen for cryptic species within the A. fumigatus complex (e.g., A. lentulus), all A. fumigatus isolates for which the MIC of any azole was ≥2 μg/ml were tested for growth at 50°C. All isolates screened grew at 50°C, confirming that they were likely to be A. fumigatus.
Antifungal susceptibility testing.Itraconazole (Janssen), posaconazole (Schering-Plough), and voriconazole (Pfizer) were all obtained as reagent-grade powders from their respective manufacturers. The BMD method was performed according to the CLSI M38-A2 standard (7). Trays containing a 0.1-ml aliquot of the appropriate drug solution (2× the final drug concentration) in each well were sealed and stored at −70°C until used in the study. The stock conidial suspension (106 spores/ml) was diluted to a final inoculum concentration of 0.4 × 104 to 5 × 104 CFU/ml and dispensed into the microdilution wells. The final concentrations of drugs in the wells ranged from 0.007 to 8 μg/ml. The inoculated microdilution trays were incubated at 35°C and read at 48 h. The MIC endpoint for the azoles was defined as the lowest concentration that produced complete inhibition of growth.
In addition, 43 isolates of A. fumigatus which had been submitted to the MEFTL as part of global antifungal surveillance projects since January 2000 and for which itraconazole MICs were ≥2 μg/ml (2 to >8 μg/ml) were used to assess cross-resistance patterns.
Quality control.Quality control was ensured by testing the following strains recommended in CLSI standard M38-A2 (7): Candida parapsilosis ATCC 22019, Candida krusei ATCC 6258, and Aspergillus flavus ATCC 204304.
Definitions.The definitions of WT and ECVs were those outlined previously by Kahlmeter et al. (25), Turnidge and Paterson (46), and Rodriguez-Tudela et al. (44). A WT organism is defined as a strain which does not harbor any acquired resistance to the particular antimicrobial agent (azole) being examined (46). In this particular instance, a WT MIC distribution for the azoles and A. fumigatus was ensured by omitting those isolates for which an itraconazole MIC of 4 μg/ml or greater was obtained (44).
The ECV for each azole was obtained by considering the WT MIC distribution, the modal MIC for each distribution, and the inherent variability of the test (usually ±1 log2 dilution). In general, the ECV should encompass at least 95% of isolates in the WT distribution (43). Organisms with acquired resistance mechanisms may be identified as those with a MIC higher than the highest MIC of the WT (>ECV) (25, 44).
RESULTS AND DISCUSSION
The WT MIC distributions for A. fumigatus and the three triazoles are shown in Fig. 1. These distributions are essentially the same as those described by Rodriguez-Tudela et al. (44) using the EUCAST BMD method. By excluding those isolates for which the itraconazole MICs were ≥4 μg/ml, we have ensured that the remaining isolates comprising these MIC distributions are free of acquired azole resistance mutations (44). The modal MICs for itraconazole, posaconazole, and voriconazole were, respectively, 0.25 μg/ml (41.3% of all values), 0.03 μg/ml (43.2% of all values), and 0.25 μg/ml (53.5% of all values) (Table 1 and Fig. 1). The modal MIC ± 1 twofold dilution comprised 90.3% of the strains for itraconazole, 88.9% of the strains for posaconazole, and 94.0% of the strains for voriconazole. The limitations of the twofold dilution scheme used in the CLSI BMD method are especially apparent for itraconazole, where the true mode lies between 0.12 and 0.25 μg/ml. Similarly, the MIC distributions for both posaconazole and voriconazole show a prominent shoulder at the next higher MIC immediately adjacent to the designated modal MIC (encompassing 29.7% of the posaconazole MICs and 38.6% of the voriconazole MICs) (Fig. 1). Given these rather broad central tendencies, visual inspection of the MIC distributions supports the ECVs assigned by Rodriguez-Tudela et al. (44): ≤1 μg/ml for itraconazole and voriconazole and ≤0.25 μg/ml for posaconazole (Table 1). These cutoffs comprise 99.8% of itraconazole and posaconazole MICs and 99.2% of voriconazole MICs.
MIC distributions for itraconazole (A), posaconazole (B), and voriconazole (C) for 637 isolates of A. fumigatus. Only isolates for which itraconazole MICs were ≤2 μg/ml were included.
MIC distribution and ECVs for azoles and A. fumigatus
Given these findings, isolates for which the itraconazole MICs are 2 μg/ml or greater may be considered as outliers with an increased probability of harboring an azole resistance mechanism (44). Table 2 shows the results of cross-resistance studies using a collection of 43 isolates of A. fumigatus for which itraconazole MICs fell outside of the ECV: range of 2 to >8 μg/ml. A high degree of cross-resistance is apparent between itraconazole and posaconazole but not between itraconazole and voriconazole. The posaconazole MIC distribution for these isolates is clearly different from the WT distribution, whereas that for voriconazole is essentially the same as the WT distribution (Tables 1 and 2). Although the mechanisms of resistance were not determined in this study, the itraconazole-posaconazole cross-resistance phenotype observed here suggests a G54 substitution in cyp51A (44). It has been postulated, based on molecular modeling studies, that a G54 substitution confers resistance to posaconazole and itraconazole by perturbing the binding of the long side chain in the hydrophobic channel of the enzyme (50). Substitutions at G54 would be predicted to have less effect on the binding of voriconazole, which lacks a long side chain. The MICs for three of the isolates were higher than the ECVs for all three azoles, suggesting either an M220 substitution or the L98H TR.
Cross-resistance between itraconazole, posaconazole, and voriconazole among 43 isolates of A. fumigatus with decreased susceptibility to itraconazole
The results of this study confirm those of Rodriguez-Tudela et al. (44) and demonstrate that the WT MIC distribution and ECVs for A. fumigatus and the mold-active triazoles are essentially the same when determined by either the CLSI or EUCAST BMD method. Although azole resistance is rare among clinical isolates of A. fumigatus, it does appear to be increasing and thus warrants continued surveillance (2, 6, 9, 23, 47). We agree with Rodriguez-Tudela et al. (44) that isolates of A. fumigatus that are susceptible to itraconazole thus far have not displayed resistance to other azole antifungal agents, and this agent will continue to prove useful in tracking the emergence of azole resistance and cross-resistance profiles of A. fumigatus. Although ECVs do not take the place of clinical breakpoints, they will be very useful in resistance surveillance and serve as an important step in the establishment of clinical breakpoints.
ACKNOWLEDGMENTS
We thank Caitlin Howard for excellent secretarial support.
This study was supported in part by research and educational grants from Pfizer Pharmaceuticals and the Schering-Plough Research Institute.
FOOTNOTES
- Received 12 May 2009.
- Returned for modification 29 July 2009.
- Accepted 9 August 2009.
- Copyright © 2009 American Society for Microbiology