Multicenter Study of Isavuconazole MIC Distributions and Epidemiological Cutoff Values for Aspergillus spp. for the CLSI M38-A2 Broth Microdilution Method

Abstract

Epidemiological cutoff values (ECVs) were established for the new triazole isavuconazole and Aspergillus species wild-type (WT) MIC distributions (organisms in a species-drug combination with no detectable acquired resistance mechanisms) that were defined with 855 Aspergillus fumigatus, 444 A. flavus, 106 A. nidulans, 207 A. niger, 384 A. terreus, and 75 A. versicolor species complex isolates; 22 Aspergillus section Usti isolates were also included. CLSI broth microdilution MIC data gathered in Europe, India, Mexico, and the United States were aggregated to statistically define ECVs. ECVs were 1 μg/ml for the A. fumigatus species complex, 1 μg/ml for the A. flavus species complex, 0.25 μg/ml for the A. nidulans species complex, 4 μg/ml for the A. niger species complex, 1 μg/ml for the A. terreus species complex, and 1 μg/ml for the A. versicolor species complex; due to the small number of isolates, an ECV was not proposed for Aspergillus section Usti. These ECVs may aid in detecting non-WT isolates with reduced susceptibility to isavuconazole due to cyp51A (an A. fumigatus species complex resistance mechanism among the triazoles) or other mutations.

INTRODUCTION

Fungal infections caused by the Aspergillus fumigatus species complex and other Aspergillus spp. are common and are usually associated with high morbidity and mortality rates, especially in immunocompromised hosts (1–4). The triazoles itraconazole, voriconazole, and posaconazole have a broad spectrum of in vitro activity against molds and are important therapeutic agents for the systemic treatment and prevention of aspergillosis (5). Isavuconazole (BAL4815; codeveloped by Basilea Pharmaceutica International Ltd., Basel, Switzerland, and Astellas Pharma, Tokyo, Japan) is a newer water-soluble triazole, with favorable pharmacodynamic and pharmacokinetic (PK/PD) parameters, that is under clinical evaluation (phase III) for the treatment of invasive aspergillosis and candidiasis (6–8). Similar to the other azoles, isavuconazole's mode of action is the inhibition of ergosterol biosynthesis (the enzymes 14-α-sterol demethylases A and B, which are encoded by the cyp51A and cyp51B genes, respectively). Isavuconazole has in vitro and in vivo activities similar to those of licensed triazoles against Aspergillus spp.; other molds, including the mucormycetes (zygomycetes); as well as Candida spp. (9–13). Acquired azole resistance in Aspergillus spp., associated with mutations of the cyp51 gene, has been documented since the1990s, and it appears to have increased in recent years, especially in Europe (14–20).

Although individual laboratories have evaluated the in vitro activity of isavuconazole against both yeasts and filamentous fungi by Clinical and Laboratory Standards Institute (CLSI) and European Committee on Antimicrobial Susceptibility Testing (EUCAST) methodologies, epidemiological cutoff values (ECVs), based on MIC data from multiple laboratories (at least three laboratories) as well as different geographical areas, have not been established for this agent and Aspergillus spp. Because of this, we have defined isavuconazole ECVs for four of the six species complexes evaluated (A. fumigatus, A. flavus, A. niger, and A. terreus) and tentative susceptibility cutoffs for the A. nidulans species complex and the A. versicolor species complex. Sufficient MICs for the latter two species were not available to calculate a more definite ECV. Wild-type (WT) distributions of each species were calculated by using aggregated MIC data gathered from three to eight laboratories in Europe, India, Mexico, and the United States (75 to 855 MICs according to species); ECVs were not defined for Aspergillus section Usti (22 isolates comprising four species), because at least 100 MIC data points originating from three or more laboratories are recommended (CLSI Establishing ECOFFs Workshop, Tampa, FL, January 2013).

MATERIALS AND METHODS

Isolates.

Each isolate was recovered from unique clinical specimens at the following medical centers: VCU Medical Center, Richmond, VA; Vallabhbhai Patel Chest Institute, University of Delhi, Delhi, India; Universidad Autónoma de Nuevo León, Monterrey, Nuevo León, México; The Innsbruck Medical University, Innsbruck, Austria; Hospital Universitario de Valme, Seville, Spain; Canisius Wilhelmina Hospital, Nijmegen, Netherlands; Hospital General Universitario Gregorio Marañón, Faculty of Medicine, Universidad Complutense, Madrid, Spain; and JMI Laboratories, North Liberty, IA. The total aggregated maximum available MIC data for each species complex were obtained for 855 isolates of A. fumigatus, 444 of A. flavus, 106 of A. nidulans, 207 of A. niger, 384 of A. terreus, 22 of Aspergillus section Usti (comprising A. calidoustus [17 isolates], A. pseudodeflectus [2 isolates], A. ustus [2 isolates], and A. insuetus [1 isolate]) (T. Peláez, P. Escribano, C. Padilla, B. Gama, J. Guinea, A. Espinel-Ingroff, and E. Bouza, unpublished data), and 75 of A. versicolor. Identification of most of the isolates to the species level was performed in each laboratory by using conventional methods (both macroscopic and microscopic characteristics on Sabouraud and Czapek-Dox agars) (21), but the species listed for Aspergillus section Usti were identified by partially amplifying and sequencing the β-tubulin (benA) gene, using βtub1/2 primers (22, 23; Peláez et al., unpublished). We had at least six A. fumigatus species complex isolates with documented acquired resistance to triazoles (MICs of itraconazole, voriconazole, and isavuconazole of ≥4 μg/ml) and confirmed mechanisms of resistance (14, 17). Data for at least one of four quality control (QC) isolates, Candida parapsilosis ATCC 22019, Candida krusei ATCC 6258, Paecilomyces variotii ATCC MYA-3630, and A. flavus ATCC 204394 (24, 25), were reported by the participant laboratories.

Antifungal susceptibility testing.

Isavuconazole MIC results for each available isolate in the total set (Tables 1 and 2) were obtained by each center according to the CLSI M38-A2 broth microdilution method (standard RPMI 1640 broth [0.2% dextrose] and final inoculum concentrations that ranged from 0.4 × 10⁴ to 5 × 10⁴ CFU/ml); MICs were the lowest drug concentrations that produced complete growth inhibition (100%) at 48 h. MICs of licensed triazoles for the isolates for which isavuconazole MICs were high (>4 μg/ml) were determined in the same manner. Isavuconazole MICs for the QC Candida strains were obtained after 48 h by using 50% growth inhibition criteria. These are the optimal testing conditions recently identified for isavuconazole and Aspergillus spp. and for Candida spp. after 48 h of incubation (24, 25).

Table 1.

Pooled MIC distributions of isavuconazole for Aspergillus spp. from two to eight laboratories, using the CLSI M38-A2 microdilution method^a

Species complex or section	No. of isolates/no. of laboratories	No. of isolates with MIC (μg/ml) of:
		0.03	0.06	0.125	0.25	0.5	1.0	2.0	4.0	≥8.0
A. fumigatus	855/8		6	31	149	508	113	33	4	11
A. flavus	444/7			2	29	253	146	13	1
A. nidulans	106/3		7	51	19	17	12
A. niger	207/6		1	4	11	52	75	55	7	2
A. terreus	384/4		4	32	171	162	14	1
A. versicolor	75/3	5	3	20	24	17	4	1		1
Aspergillus section Ustib	22/2				2		10	10

^aWild-type MIC distributions for each species were obtained by using the CLSI broth microdilution method (M38-A2) and according to the recently identified testing conditions for Aspergillus spp. and isavuconazole: 48 h of incubation and 100% growth inhibition (24, 25). Values in boldface type indicate modes or most frequent MICs for each species. The following are species complexes, as per nonmolecular identification (21): A. fumigatus, A. flavus, A. nidulans, A. niger, A. terreus, and A. versicolor.

^bAspergillus section Usti comprised the following species, as per molecular identification: A. calidoustus (17 isolates), A. pseudodeflectus (2 isolates), A. ustus (2 isolates), and A. insuetus (1 isolate) (22, 23; Peláez et al., unpublished).

Table 2.

Isavuconazole ECVs based on MICs from three to eight laboratories and as determined by the CLSI M38-A2 broth microdilution method

Species complex or sectiona	No. of isolates/no. of laboratories	MIC range (μg/ml)	Mode (μg/ml)b	ECVc			% observed above ECV
				95%	97.5%	99%	ECV 95%	ECV 97.5%	ECV 99%
A. fumigatus	855/8	0.06–8	0.5	1	1	1	5.6	5.6	5.6
A. flavus	444/7	0.06–2	0.5	1	1	2	3.2	3.2	0.2
A. nidulansd	106/3	0.06–1	0.12	0.25	0.25	0.25	27.4	27.4	27.4
A. niger	207/6	0.06–>8	1	4	4	4	1.0	1.0	1.0
A. terreus	386/5	0.06–2	0.25	1	1	1	0.3	0.3	0.3
A. versicolord	75/3	0.03–>8	0.25	1	1	2	2.7	2.7	1.3
Aspergillus section Ustie	22/2	0.25–2	1	ND	ND	ND	ND	ND	ND

^aThe following are species complexes, as per nonmolecular identification (21): A. fumigatus, A. flavus, A. nidulans, A. niger, A. terreus, and A. versicolor. Aspergillus section Usti comprised the following species, as per molecular identification: A. calidoustus (17 isolates), A. pseudodeflectus (2 isolates), A. ustus (2 isolates), and A. insuetus (1 isolate) (22, 23; Peláez et al., unpublished).

^bMost frequent MIC.

^cCalculated ECVs comprising ≥95, ≥97.5, or ≥99% of the statistically modeled MIC population.

^dTentative values until more data are obtained.

^eND, not determined due to insufficient numbers of laboratories and isolates/species.

Definitions.

The wild type (WT) is defined as the population of strains in a species-drug combination with no detectable acquired resistance mechanisms, and the ECV is the highest WT MIC. The ECV categorizes an isolate as either WT or non-WT, and it has been described as either the WT cutoff value (CO_WT) or the epidemiological cutoff value (ECOFF or ECV) (26).

We use the ECV term throughout the present report because it has been previously used in similar fungal studies (26–29).

Data analysis.

The MIC distributions of each of the seven species tested in each participant laboratory were first screened for evidence of grossly abnormal distributions, and modal (most frequent) MICs for each laboratory were determined (28, 29). Grossly skewed distributions and distributions which had a modal MIC at the lowest concentration tested were excluded. Next, the WT distributions were obtained by pooling qualifying MIC distributions from participant laboratories for each species, and the ECV was then estimated by statistical methods (30). A minimum of 3 laboratories and 100 data points were required to establish a reasonable estimated ECV for a species. In the statistical method, the modeled WT population is based on fitting a normal distribution at the lower end of the MIC range, calculating the mean and standard deviation of that normal distribution, and using those values to estimate the MIC that captured at least 95%, 97.5%, and 99% of the modeled WT population, rounded up to the nearest 2-fold dilution. The modes for each species, the inherent variability (±1 log₂ dilution) of susceptibility testing, as well as a search for outlier laboratories in each distribution were also considered (31).

RESULTS AND DISCUSSION

For the MIC result to be clinically useful, it should categorize the isolate as either susceptible (treatable), intermediate (possibly treatable), or resistant (nontreatable) to the specific agent; this result is the clinical breakpoint (CBP) or the predictor of clinical outcome (26, 32). CBPs are established based on clinical trial data, global surveillance and ECVs, resistance mechanisms, and PK/PD parameters from model systems. The CLSI has not defined CBPs for molds due to the insufficient number of resistant isolates recovered during clinical trials; most isolates are WT strains. ECVs for Aspergillus spp. and the licensed triazoles (itraconazole, posaconazole, and voriconazole), caspofungin, and amphotericin B were recently defined according to current criteria for ECV definition (CLSI data from multiple laboratories [at least three] and for ≥100 isolates). In the present study, ECVs of isavuconazole for four Aspergillus species complexes as well as tentative values for two other species were established. It is expected that these susceptibility cutoffs may aid in the evaluation of clinical isolates by detecting those strains with reduced isavuconazole susceptibility due to cyp51 mutations and may serve as an early warning of emerging changes in the susceptibility patterns of these organisms.

Despite standardized testing conditions, variability is usually evident during comparison of MIC results from multiple laboratories, including studies where single strains are evaluated in order to select QC isolates (24, 25). The performance of antimicrobial susceptibility testing is monitored by the introduction of at least one QC or control strain for which MIC ranges of acceptable reproducibility (within 3 to 4 dilutions) have been established. Recent collaborative studies have identified the optimum parameters for testing isavuconazole and Aspergillus spp., and two yeasts and two molds were selected as QC strains and MIC limits for these isolates were proposed (25). In the present study, each participating laboratory reported isavuconazole MIC data for at least one of these four QC isolates in the following ranges: 0.015 to 0.06 μg/ml in three laboratories for C. parapsilosis ATCC 22019, 0.12 to 0.5 μg/ml in four laboratories for C. krusei ATCC 6258, 0.5 to 2 μg/ml in four laboratories for A. flavus ATCC 204304, and 0.06 to 0.5 μg/ml in two laboratories and 0.12 to 2 μg/ml in another laboratory for Paecilomyces variotii ATCC 3630. Therefore, most values were within the 3- to 4-dilution MIC range for the QC strains (25). Reproducibility was similar to that observed in the collaborative study proposing isavuconazole QC MIC ranges.

Table 1 depicts isavuconazole aggregated MIC distributions for the seven Aspergillus spp. The graphs indicated that the distributions of most species were symmetrical; the exception was the MIC distribution for the A. nidulans species complex. The skewed distribution of the A. nidulans species complex may have been due to the smaller numbers of isolates and laboratories (three laboratories and 106 isolates [Table 1]). It is noteworthy that most modal MICs for individual contributing laboratories for all the species evaluated were equal to or within one 2-fold dilution, including the modes (0.12 and 0.25 μg/ml) from the three laboratories that provided MICs for the A. nidulans species complex.

Table 2 depicts the proposed isavuconazole ECVs (using ≥95%, ≥97.5%, and ≥99% of the modeled MIC population) as well as the range and modal MICs for each of the seven Aspergillus spp. The lowest modal MIC (0.12 μg/ml) was for the A. nidulans species complex, and the highest was for the A. niger species complex and Aspergillus section Usti (1 μg/ml). Data from a previous study in which ECVs for Aspergillus spp. and the three other triazoles were defined allowed for the following comparisons between the data sets (27). The isavuconazole modal MIC for the A. fumigatus species complex was higher than those of posaconazole and voriconazole (0.5 μg/ml versus 0.06 and 0.25 μg/ml, respectively) but was the same as that of itraconazole; a similar pattern was evident for the A. flavus species complex. The isavuconazole mode for the A. terreus species complex was the same as those of itraconazole and posaconazole (0.25 μg/ml); a mode of 0.5 μg/ml has been reported for voriconazole. As for itraconazole, the isavuconazole mode for the A. niger species complex was higher than those for the other common species (1 μg/ml versus 0.25 to 0.5 μg/ml) as well as those for the A. nidulans species complex and the A. versicolor species complex (0.12 and 0.25 μg/ml, respectively). To our knowledge, aggregated MIC data from multiple laboratories have not been published for the Aspergillus section Usti group of species to allow modal MIC comparisons, but high triazole MICs have been reported for this group (33). Most isavuconazole modal MICs ±1 2-fold dilution comprised similar percentages of the populations for five of the seven species: 87.9% of A. niger species complex, 89.7% of Aspergillus section Usti, 90% of A. fumigatus species complex, 95.1% of A. terreus species complex, and 96.4% of A. flavus species complex isolates. Percentages were lower for A. versicolor species complex (81.3%) and A. nidulans species complex (72.6%) isolates. Since we have aggregated MIC data for our definition of isavuconazole ECVs from a wide geographical range, as in previous ECV studies (27–29), and little modal variation was observed among the laboratories contributing the MIC data for the majority of the species evaluated, we are confident in the validity of the MIC data in the present study.

The isavuconazole ECV for the A. fumigatus species complex, the A. flavus species complex, and the A. terreus species complex, encompassing ≥95% of the modeled MIC population, was 1 μg/ml; ECVs encompassing either ≥97.5% or ≥99% of the population were either the same or 1 dilution higher (Table 2). An ECV of 1 μg/ml has been reported for these three species versus itraconazole and voriconazole, while posaconazole ECVs have been consistently lower (0.12 to 0.5 μg/ml) using both CLSI and EUCAST methods as well as other statistical approaches (27, 34–36). Of the 855 isolates, 5.6% had MICs exceeding the estimated ECV of 1 μg/ml, and two of the eight laboratories that contributed data had modal MICs at the ECV. Although cyp51 gene mutations have been found in isolates with isavuconazole MICs of >2 μg/ml (14, 17), there is no genetic information for isolates for which MICs are 2 μg/ml. Such information and more MIC data would corroborate our estimated ECV for this species complex. In the meantime, the estimated ECV of 1 μg/ml would avoid misclassification of non-WT isolates as WT isolates. The frequency of isavuconazole MICs of >1 μg/ml (non-WT isolates) for A. fumigatus species complex isolates was 5.6% (48 isolates) (Table 2), which represents a higher rate than previously described for this species with the other three triazoles (2.2 to 3.1%) (27). However, we have a lower sample size in the present study (855 versus >1,600 isolates). Among the 855 A. fumigatus species complex isolates, 15 isolates recovered from Europe (10 isolates), India (3 isolates), Mexico (1 isolate), and the United States (1 isolate) had isavuconazole MICs of ≥4 μg/ml. Twelve of the 15 isolates contained cyp51A with a TR₃₄ promoter duplication and an L98H mutation that correlated with phenotypic triazole cross-resistance (Table 3) (14, 17). Modal MICs for these isolates are above ECVs defined for the A. fumigatus species complex and the licensed triazoles (27, 34, 35). Correlation between triazole MICs above the ECV (≥4 μg/ml) in the A. fumigatus species complex, single or multiple point mutations, and patient failure on licensed triazole treatment has been reported (14–20).

Table 3.

Cross-resistance between isavuconazole (MICs of ≥4 μg/ml) and three licensed triazoles among 26 Aspergillus isolates, as determined by the CLSI M38-A2 microdilution method at 48 h^a

Species complex (no. of isolates)	Range (mode)b (μg/ml) withd:
	Itraconazole	Posaconazole	Voriconazole	Isavuconazole
A. fumigatus (15)c	1–≥16 (≥16)	0.5–8 (2)	0.5–8 (8)	4–8 (8)
A. flavus (1)	1 (NA)	1 (NA)	1 (NA)	4 (NA)
A. niger (9)	1–16 (2, 8)	0.5–1 (1)	1–≥16 (2)	4–≥16 (4)
A. versicolor (1)	8 (NA)	1 (NA)	4 (NA)	>8 (NA)

^aIsavuconazole MICs of ≥4 μg/ml were not reported for the other species evaluated. Itraconazole, posaconazole, and voriconazole ECVs for the A. fumigatus species complex and the A. flavus species complex were 1, 0.25, and 1 μg/ml, respectively; those for the A. niger species complex were 2, 0.5, and 2 μg/ml, respectively; and those for the A. versicolor species complex were 2, 1, and 2 μg/ml, respectively (27).

^bMost frequent MIC.

^cTwelve of the 15 isolates contained a TR₃₄ promoter duplication and an L98H mutation in the cyp51A gene (14, 17).

^dNA, not applicable, because modes were not determined for single isolates.

To our knowledge, there is only one report of triazole ECVs for non-A. fumigatus species (27); our comparison of non-WT rates for these species are based on results from that study. There were 14 isavuconazole MICs above the ECV for non-WT isolates for the A. flavus species complex (MICs of >1 μg/ml; 3.2%) (Tables 1 and 2); percentages of non-WT isolates have been variable for this species and the licensed triazoles (posaconazole, 5.6%; voriconazole, 2%; itraconazole, 0.7%). There were only a few isavuconazole non-WT isolates for the A. terreus species complex (MIC of >1 μg/ml; 0.3%); these results reflect the rate of non-WT isolates found among itraconazole and posaconazole distributions for this species, while the rate of isavuconazole non-WT values was lower than that for voriconazole (0.3 and 3%, respectively) (27). Our isavuconazole ECV for the A. niger species complex (an Aspergillus section Nigri group member) was 2 dilutions higher (4 μg/ml) than those for the other three more common species (Table 2); both itraconazole and voriconazole ECVs for this species have been higher than that for posaconazole (2 and 0.5 μg/ml, respectively) (27). There were only two isavuconazole non-WT isolates of the A. niger species complex (MIC > 4 μg/ml) (Table 1); however, cross-resistance was observed with both itraconazole and voriconazole for one isolate (Table 3). In general, isavuconazole MIC₉₀s for A. niger species complex isolates have been consistently 1 dilution higher than those for other Aspergillus spp. (7, 10), and reduced susceptibility to licensed triazoles (most itraconazole MICs, ≥4 μg/ml; voriconazole MICs, 2 μg/ml; posaconazole MICs, 0.12 to 0.5 μg/ml) have been documented in Europe for A. tubingensis and A. foetidus (Aspergillus section Nigri group members) (37, 38). These high triazole MICs are a concern because the A. niger species complex is more frequently involved in pulmonary aspergillosis. Molecular resistance mechanisms have not been identified for the A. niger species complex, but azole resistance was correlated with the multiplication of cyp51A in an engineered laboratory strain of this species (38–40). Two cyp51 genes were reported in A. terreus species complex isolates, and more recently, a Cyp51Ap M217I alteration was found in isolates with elevated EUCAST itraconazole MICs (1 to 2 μg/ml), but such MICs were also found in the absence of alterations; isavuconazole was not evaluated in that study (41). Although the clinical relevance of mold testing remains uncertain in the absence of breakpoints, both CLSI and EUCAST methodologies can identify emerging triazole resistance (MICs above the ECV).

The number of isolates for which isavuconazole MICs were available was substantially smaller for the A. nidulans species complex, the A. versicolor species complex, and Aspergillus section Usti (Tables 1 and 2), because these species are less frequently associated with disease. Therefore, the required total numbers of isolates and laboratories allowed the definition of only a tentative ECV of 1 μg/ml for the A. versicolor species complex, with 2.7% of the isolates being above this tentative cutoff; cross-resistance was observed for one isolate (Table 3). Although an isavuconazole ECV for the Aspergillus section Usti members could not be proposed because this distribution contained several species and the number of isolates is small, the MIC distribution and mode, as listed in Tables 1 and 2, are important for future studies, since little information is available in the literature regarding the patterns of susceptibility of this fungal group to antifungal agents (33, 42). While the distribution of A. nidulans species complex isolates fulfilled the criteria for ECV definition, we are proposing a tentative isavuconazole ECV of 0.25 μg/ml, because as mentioned above, its MIC distribution was skewed. Therefore, the percentage of MICs above the ECV (27.4%) could be erroneously high (Table 2). This tentative ECV may change when more laboratories and MICs become available. Mechanisms of resistance have not been documented for these less common species, despite the reported reduced susceptibility of A. versicolor species complex isolates to the three licensed triazoles (38). Amplification or overexpression of the cyp51 gene in an A. nidulans species complex laboratory mutant has been correlated with triazole resistance (43), which suggested that mechanisms of resistance in the non-A. fumigatus species could be similar to those in the A. fumigatus species complex.

In conclusion, data originating from three to eight laboratories enable us to propose species-specific isavuconazole ECVs of 1 μg/ml for the A. fumigatus species complex, the A. flavus species complex, and the A. terreus species complex and 4 μg/ml for the A. niger species complex. In addition, we have proposed tentative isavuconazole ECVs for the A. nidulans species complex (0.25 μg/ml) and the A. versicolor species complex (1 μg/ml) and have provided a susceptibility MIC distribution for Aspergillus section Usti. The availability of CLSI standard parameters for testing of Aspergillus with isavuconazole in addition to our ECVs would aid in monitoring emerging isavuconazole resistance in Aspergillus spp. as well as to distinguish non-WT from WT isolates.