Introduction

The systemically active antifungal armamentarium currently includes the polyenes, flucytosine, fluconazole, the extended-spectrum triazoles (itraconazole, posaconazole, and voriconazole), and the echinocandins [1–6]. Despite the fact that in total, these agents cover the vast majority of opportunistic fungal pathogens and are increasingly employed in either a prophylactic or preemptive treatment strategy [2, 7–10], breakthrough invasive fungal infections (bIFI) continue to be reported and increasingly involve yeasts and/or molds that are relatively uncommon and tend to exhibit decreased susceptibility to the available antifungal agents [4, 7, 11–19]. These observations underscore the need for continued surveillance efforts designed to provide accurate identification and resistance detection in these emerging pathogens, as well as the development and introduction of new antifungal agents to address the present need for broad-spectrum antifungal coverage [4, 12, 20, 21]. The recent development of species-specific clinical breakpoints (CBPs) and epidemiological cutoff values (ECVs) for in vitro susceptibility testing of Candida spp. against both triazoles and echinocandins further emphasize the need for accurate identification of fungal pathogens [13, 22]. Indeed, erroneous identification of fungal species will now have potentially important effects on the interpretation of MIC data and appropriateness of therapeutic decisions [13, 22–25].

Isavuconazole is an investigational triazole that is presently in late-stage clinical development for the treatment of invasive candidiasis, invasive aspergillosis, and infections due to non-Aspergillus molds [26]. Isavuconazole may be administered orally or parenterally and exhibits broad antifungal activity against common and uncommon fungal pathogens, including Candida, Aspergillus, non-Candida yeasts, and non-Aspergillus molds [27–35].

Previously, we demonstrated a high level of concordance between isavuconazole MIC results produced with the Clinical and Laboratory Standards Institute (CLSI) and the European Committee on Antimicrobial Susceptibility Testing (EUCAST) broth microdilution (BMD) methods in testing Candida spp., suggesting that either method could be employed for the purpose of resistance surveillance [31]. In the present study, we use the database from the global SENTRY Antifungal Surveillance Program (2012) to further document the in vitro activity of isavuconazole as determined by CLSI BMD methods against both common and uncommon species of Candida causing IFIs. We compare the activity of isavuconazole with that of the established triazole and echinocandin antifungal agents and examine the extent of decreased susceptibility to isavuconazole among isolates classified as susceptible dose dependent (SDD), resistant (R), or non-wild type (non-WT) to fluconazole and voriconazole. Importantly, we employ DNA sequence analysis to confirm the identification of uncommon species of Candida to ensure an accurate assessment of the antifungal MIC phenotype for each species.

Materials and Methods

Organisms and Sources

A total of 1,421 non-duplicate strains of Candida were collected prospectively from 75 medical centers located in North America (30 sites), Europe (24 sites), Latin America (10 sites), and the Asia–Pacific region (11 sites). These strains were recovered consecutively from patients with bloodstream infections (1,094 strains), from normally sterile body fluids, tissues, and abscesses (162 strains), and 165 were collected from non-specified infection sites.

Isolates were identified at participant institutions using methods routinely employed at the submitting laboratory including the use of Vitek, MicroScan, API, and AuxaColor systems supplemented by classical methods for yeast identification [36, 37]. Isolates were submitted to JMI Laboratories (North Liberty, Iowa, USA), where the identification was confirmed by morphological, biochemical, and molecular methods [21–23, 25, 31]. Yeast isolates were subcultured and screened using CHROMagar Candida (Becton–Dickinson, Sparks, Maryland, USA) to ensure purity and to differentiate Candida albicans/Candida dubliniensis, Candida tropicalis, and Candida krusei. Biochemical and physiological tests including Vitek 2 (bioMerieux, Hazelwood, Missouri, USA) testing, trehalose assimilation (Candida glabrata), and growth at 45 °C (C. albicans/C. dubliniensis) were also used to establish the identification of common (C. albicans, C. glabrata, C. parapsilosis, C. tropicalis, and C. krusei) Candida species. Molecular methods were used for common species of Candida that could not be definitively identified using phenotypic methods or that presented unusual phenotypic or biochemical profiles, as well as all uncommon species of Candida. Candida spp. were identified using sequence-based methods for the internal transcribed spacer (ITS) region and 28S ribosomal subunit (D1/D2) [22, 23, 25]. Nucleotide sequences were examined using Lasergene software (DNAStar, Madison, Wisconsin, USA) and then compared to database sequences using BLAST (http://www.nbeti.nlmnih.gov.blast). Results were considered acceptable if homology was >99.5 % with other entries in the databases used for comparison. Available sequences that were considerably different from the majority of entries for one species were considered outliers and were discarded in the analysis. Additionally, if no match was found in the database, the ID was based on species complex (SC), genus, family, or order, according to the most current classification systems.

Among the isolates of Candida (1,421), there were 671 isolates of C. albicans, 291 of C. glabrata, 236 of C. parapsilosis, 122 of C. tropicalis, 32 of C. krusei, 24 of Candida lusitaniae, 13 of C. dubliniensis, and 32 of miscellaneous Candida species (1 C. bracarensis, 1 C. catenulata, 2 C. fabianii, 2 C. fermentati, 1 C. fluviatilis, 6 Candida guilliermondii, 1 C. haemulonii, 8 C. kefyr, 2 C. lipolytica, 2 C. orthopsilosis, 2 C. pelliculosa, 1 C. rugosa, 1 C. thasaenensis, 1 C. thermophila, and 1 Candida spp. not further identified).

In addition to these isolates, we included additional isolates of less-common species from the SENTRY Program (2011) database: 5 C. fermentati, 8 C. guilliermondii, 10 C. kefyr, and 2 C. lipolytica. The MIC results for these isolates were added to those from 2012 to augment the numbers of these unusual species.

Antifungal Susceptibility Testing

All yeast isolates were tested for in vitro susceptibility to amphotericin B, flucytosine, the triazoles (fluconazole, isavuconazole, itraconazole, posaconazole, and voriconazole), and the echinocandins (anidulafungin, caspofungin, and micafungin) using CLSI [38] BMD methods. The MIC results for all agents were read following 24 h of incubation at 35°C. MIC values were determined visually, as the lowest concentration of drug that caused complete (amphotericin B) or significant (≥50 %) growth diminution levels (all other agents) [38, 39].

We used the recently revised CLSI clinical breakpoint (CBP) values to identify strains of the five most common species of Candida (C. albicans, C. glabrata, C. parapsilosis, C. tropicalis, and C. krusei) that were susceptible and resistant to the echinocandins as well as those that were susceptible and resistant to fluconazole and voriconazole [22, 39]: anidulafungin, caspofungin, and micafungin MIC values of ≤0.25 and ≥1 µg/ml were categorized as susceptible and resistant, respectively, for C. albicans, C. tropicalis, and C. krusei, and MIC results of ≤2 and ≥8 µg/ml were categorized as susceptible and resistant, respectively, for C. parapsilosis; anidulafungin and caspofungin MIC values of ≤0.12 and ≥0.5 µg/ml and micafungin MIC values of ≤0.06 and ≥0.25 µg/ml were considered susceptible and resistant, respectively, for C. glabrata; fluconazole MIC results of ≤2 and ≥8 µg/ml were defined as susceptible and resistant, respectively, for C. albicans, C. parapsilosis, and C. tropicalis, and MICs of ≤32 and ≥64 µg/ml were considered susceptible dose dependent (SDD) and resistant, respectively, for C. glabrata. All isolates of C. krusei were defined as resistant to fluconazole. The CLSI susceptible and resistant breakpoints for voriconazole are ≤0.12 and ≥1 µg/ml, respectively, for C. albicans, C. parapsilosis, and C. tropicalis, and MIC results of ≤0.5 and ≥2 µg/ml were categorized as susceptible and resistant, respectively, for C. krusei; CLSI has not assigned CBPs for voriconazole and C. glabrata and recommends the ECV of 0.5 µg/ml to be used to differentiate wild type (WT; MIC ≤ ECV) from non-WT (MIC > ECV) strains of this species [22, 39].

CBPs have not been established for amphotericin B, flucytosine, itraconazole, isavuconazole, or posaconazole and for any antifungal agent and the less-common species of Candida; however, ECVs have been established for amphotericin B, flucytosine, the triazoles (fluconazole, itraconazole, posaconazole, and voriconazole), and the echinocandins and six species of Candida that are encountered less frequently (C. lusitaniae, C. guilliermondii, C. dubliniensis, C. kefyr, C. orthopsilosis, and C. pelliculosa) [22].

Quality control was performed as recommended in CLSI document M27-A3 [38] using strains C. krusei ATCC 6258 and C. parapsilosis ATCC 22019.

Results and Discussion

Among the 21 species of Candida encountered in this survey, C. albicans was most common in Europe (56.2 %) and least common in North America (39.5 %), whereas C. glabrata was most common in North America (28.5 %) and least common in Latin America (6.9 %). C. parapsilosis was most common in Latin America (27.1 %), and C. tropicalis was most common in the Asia–Pacific region (17.2 %). C. krusei was most common in North America (2.7 %), as were other miscellaneous species of Candida (5.8 %) (data not shown).

Table 1 shows the MIC distributions for isavuconazole and the 11 most common species of Candida referred from the 75 participating centers. Among these isolates, isavuconazole was most active against C. albicans (modal MIC, 0.008 µg/ml) and least active against C. glabrata, C. krusei, and C. guilliermondii (modal MICs, 0.5 µg/ml). The vast majority of each species, with the exception of C. glabrata, C. krusei, and C. guilliermondii, were inhibited by ≤0.12 µg/ml of isavuconazole [99.0 %; range 94.3 % (C. tropicalis) to 100.0 % (C. lusitaniae and C. dubliniensis)]. C. glabrata, C. krusei, and C. guilliermondii isolates were largely susceptible to isavuconazole at MIC values ≤1 µg/ml (89.7, 96.9, and 92.8 %, respectively).

Table 1 Antifungal activity of isavuconazole against Candida spp. tested as part of the 2012 international surveillance program
Full size table

The antifungal activity of isavuconazole and nine comparator antifungal agents against 1,352 isolates of Candida spp. as determined with CLSI BMD methods are shown in Table 2. The results are categorized using CLSI CBPs and/or ECVs as appropriate. The vast majority of these isolates represented WT strains as determined by the respective ECVs and very few (C. glabrata and C. krusei) were resistant to triazoles or echinocandins based on CBPs.

Table 2 Antifungal activity of isavuconazole and comparator antifungal agents against key organism species/groups tested as part of the 2012 international surveillance program
Full size table

Among the azole antifungal agents, it is important to assess the issue of cross-resistance given the ubiquitous use of this antifungal class in various medical practice settings [3]. CBPs for isavuconazole and Candida have not yet been established; however, given that this agent is of the triazole class, it is reasonable to consider that it may be a substrate for CDR efflux pumps and also may be affected by quantitative or qualitative alterations in the lanosterol demethylase target enzyme [12, 20, 42, 44, 45]. Overexpression of cdr1 and cdr2 has been documented as the primary mechanism of resistance for C. glabrata versus fluconazole, itraconazole, posaconazole, and voriconazole [20, 46, 47]. As such, it is highly likely that isavuconazole can be pumped out of the cell as are the other members of this class. Thus, the discussion of potential cross-resistance is consistent with the available data. Having said this, it is true there are no CBPs to designate resistance to isavuconazole and so the discussion will focus on decreased susceptibility that may suggest cross-resistance, the impact of which must await data from ongoing clinical trials.

Previously, we examined cross-resistance among the triazoles (fluconazole, posaconazole, and voriconazole) using a large collection of Candida spp. from a global surveillance program [40, 41]. These studies provided strong support for the concerns of several investigators regarding the issue of cross-resistance among these agents that share both a common mechanism of action and certain mechanisms of resistance [7, 40–45]. Most importantly, these studies focused attention on C. glabrata as the species most likely to demonstrate cross-resistance among the three triazoles by virtue of expression of CDR efflux pumps as the primary mechanism of azole resistance [46, 47]. In the present study, we examine the potential for cross-resistance between fluconazole and isavuconazole (Table 3), and between voriconazole and isavuconazole (Table 4) using more than 1,300 isolates of Candida spp. from the 2012 surveillance study. One of the limitations of this analysis stems from the fact that the surveillance protocol specifies that only incident isolates be submitted for the study. Whereas this ensures that the population of each species in the surveillance collection represents a WT population, it also reduces the number of strains with resistance or decreased susceptibility to the various antifungal agents. Given this limitation, it is clear from the data in Tables 3 and 4 that isolates classified as susceptible to either fluconazole or voriconazole, with the exception of C. glabrata and C. krusei, represent WT strains with respect to isavuconazole and show MIC values that are almost all (99.7 %) ≤0.12 µg/ml. The higher MIC values of fluconazole and voriconazole observed with strains of C. glabrata classified as SDD or WT, respectively, are also reflected in those of isavuconazole where the modal MIC for isavuconazole was 0.5 µg/ml and 97 % of strains that were SDD to fluconazole and WT to voriconazole exhibited isavuconazole MIC values of ≤1 µg/ml. Although there were few strains of Candida classified as SDD, resistant, or non-WT to fluconazole and voriconazole, most showed elevated MIC results for isavuconazole as well. The exception to this statement was C. parapsilosis, where 98.7 % of 236 isolates were inhibited by ≤0.12 µg/ml of isavuconazole irrespective of their susceptibility to fluconazole or voriconazole. The most clear evidence of potential cross-resistance between isavuconazole and fluconazole or voriconazole may be seen with C. glabrata where the modal MIC for isavuconazole was 4 µg/ml for those strains that were either resistant or non-WT to fluconazole and voriconazole, respectively.

Table 3 In vitro activity of isavuconazole against 1,357 clinical isolates of Candida stratified by fluconazole susceptibility category
Full size table
Table 4 In vitro activity of isavuconazole against 1,389 clinical isolates of Candida stratified by voriconazole susceptibility category
Full size table

Several important observations can be made from this global survey. First, we confirm the excellent spectrum and potency of isavuconazole against Candida spp. Second, we have used the recently published CBPs and ECVs to demonstrate the generally low levels of resistance to the available antifungal agents in a large, contemporary (2012) global collection of molecularly characterized isolates. Finally, our results document decreased susceptibility to isavuconazole among isolates classified as R or non-WT to fluconazole and voriconazole, with the greatest emphasis on C. glabrata. This decreased susceptibility may suggest cross-resistance, the impact of which must be documented by data from clinical trials.