EUCAST Susceptibility Testing of Isavuconazole: MIC Data for Contemporary Clinical Mold and Yeast Isolates

Isavuconazole is the newest medical azole. We investigated EUCAST MICs for isavuconazole and seven comparators against 1,498 contemporary isolates (2016 to 2017).

I savuconazole is the newest medical azole with activity against a broad range of yeast and molds. It was licensed in Europe and the United States in 2015 by the EMA and FDA for the treatment of adults with invasive aspergillosis and also for mucormycosis although by the EMA only in patients for whom amphotericin B is inappropriate. In the same year, EUCAST clinical breakpoints were established for three Aspergillus species (A. fumigatus, A. nidulans, and A. terreus), and epidemiological cutoff values (ECOFFs) for these as well as for A. flavus and A. niger were determined (1). Isavuconazole given daily or weekly has also been found efficacious and noninferior to fluconazole for uncomplicated esophageal candidiasis in a randomized, double-blind, multicenter phase 2 trial, where Candida albicans was the most common cause of infection, accounting for The in vitro activity of isavuconazole against A. flavus, A. fumigatus, A. nidulans, and A. terreus isolates on a milligram-per-liter basis was comparable to those of voriconazole and amphotericin B and slightly lower than those of itraconazole and posaconazole (Table 2). For A. niger isolates, the isavuconazole modal MIC and MIC 50 were two 2-fold dilutions higher than those for voriconazole and itraconazole and four 2-fold dilutions higher than those for posaconazole and amphotericin B. As isavuconazole and the comparators are all regarded as valid options for A. fumigatus infections, the proportions of isolates with MICs above the ECOFF against A. fumigatus were compared for each compound as an indicator of relative coverage ( Table 2). The proportions of A. flavus and A. terreus isolates with MICs above the A. fumigatus ECOFF were lower for isavuconazole than for amphotericin B and voriconazole, whereas in contrast, the proportion of A. niger isolates less susceptible than A. fumigatus was highest for isavuconazole ( Table 2).
The vast majority of the MICs against the most susceptible Candida species fell at or below the lowest concentrations tested ( krusei. In all cases, the modal MIC and MIC 50 values set using the 11-dilution range were supported by the 8-dilution range adopted in the first part of the study period,

>4
Most prevalent Candida species C. albicans   suggesting a robust performance of EUCAST isavuconazole testing during the study period.
The rare Candida species were characterized by isavuconazole MICs spanning the entire Յ0.004to Ͼ4-mg/liter concentration range, suggesting differential susceptibilities of the involved species. The most resistant species was C. duobushaemulonii (n ϭ 1), with an MIC of Ͼ4 mg/liter, whereas the single C. auris isolate, originating from a Norwegian patient, as well as isolates belonging to C. bovina, C. famata, C. kefyr, C. metapsilosis, C. nivariensis, C. palmioleophila, C. pararugosa, C. pelliculosa, and C. utilis had isavuconazole MICs of Յ0.06 mg/liter. Similarly, the MICs for rare yeast were diverse, with Magnusiomyces capitatus being the isolate with the highest isavuconazole MIC of 2 mg/liter.
On a milligram-per-liter basis, the in vitro activity of isavuconazole was more similar to that of voriconazole against the yeast isolates than to that of fluconazole and similar to that of the echinocandins, except for C. parapsilosis, Cryptococcus, C. glabrata, and C. krusei (Table 4).
Correlation between susceptibilities to isavuconazole and comparators. Finally, the correlation between isavuconazole and comparator MICs was investigated (Table 5). A significant correlation was observed between isavuconazole and voriconazole MICs for all mold and Candida species, although it was weak (R 2 of Ͻ0.5 for Mucorales, C. krusei, and C. parapsilosis). Moreover, the correlation between isavuconazole and voriconazole was stronger than that between isavuconazole and any other comparator for all species except A. terreus and Mucorales species (the best correlation was observed for isavuconazole and posaconazole). A good and highly significant correlation was also observed between isavuconazole and fluconazole for C. albicans, C. glabrata, and C. tropicalis but not for C. parapsilosis.

DISCUSSION
This study confirmed previous findings of potent in vitro activity of isavuconazole against most human-pathogenic fungi (3-5, 9, 11). The rate of acquired resistance in Aspergillus and Candida spp. was overall low and stable (4). Exceptions were resistance rates of around 10% in A. fumigatus, A. terreus, C. glabrata, and C. tropicalis. However, resistance rates may be overestimated, as fungal isolates referred to a reference laboratory may not be representative of the general population. For the two Aspergillus species, however, the resistance was in part due to a stringent susceptibility breakpoint bisecting the wild-type populations of A. fumigatus and A. terreus, leading to a misclassification of some susceptible isolates as resistant. The ECOFF for isavuconazole against A. fumigatus is 2 mg/liter, but the clinical breakpoint established was one step lower because isolates with an MIC of 2 mg/liter may represent isolates with wild-type as well as mutant target gene sequences (1, 3). Thus, in a multicenter study, MICs straddled the ECOFF, with MICs of Ͼ2 mg/liter found in 25% of isolates harboring the M220I and M220V mutations and in 72.5% of the MIC readings of isolates withTR 34 /  (3). Hence, our data suggested that the majority of isolates with an MIC of 2 mg/liter will be true wild-type isolates, and the remaining minority will be characterized by slightly reduced isavuconazole susceptibility. Buil et al. found that the probability of target attainment for isolates with isavuconazole MICs of 2 mg/liter with the isavuconazole standard dose was ϳ75% (64% to 92%) when the 90% exposure index (EI 90 ) was used as the endpoint and that a trough level of Ն1.60 mg/liter (1.42 to 1.80 mg/liter) was the target (8).
Another recent study reported that approximately 10% of "real-life" clinical samples contained less than 1 mg/liter and another approximately 20% contained between 1 and 2 mg/liter of isavuconazole (12). Taken together, these observations support introducing an intermediate category for A. fumigatus and A. terreus isolates with an MIC of 2 mg/liter in a setting where therapeutic drug monitoring is available to confirm sufficient exposure. When the correlations between isavuconazole MICs and those for the comparators were analyzed, the strongest correlation overall was found for isavuconazole and voriconazole. Thus, a significant strong to moderate correlation was found for the four most common Aspergillus species as well as for the six most common Candida species except C. krusei and C. parapsilosis, for which the correlation was significant but weak (R 2 of 0.417 to 0.448), potentially due to the lack of isolates with acquired resistance for these two species. Thus, our results extend previous findings of a correlation between the azoles and, particularly, voriconazole and isavuconazole for A. fumigatus (7,8) and the findings that isavuconazole susceptibility was lower for Candida isolates with resistance or non-wild-type susceptibility to fluconazole and voriconazole (9). No correlation was found for isavuconazole compared to itraconazole or posaconazole for A. flavus or A. fumigatus. For A. fumigatus, this may not be surprising, as it is well acknowledged that some target gene mutations specifically affect itraconazole and posaconazole (3). For A. flavus, we assume that the absence of isolates with differential susceptibility explains the lack of correlation because MIC variation in such cases can be explained solely by test variation. Therefore, taken together, our data support that voriconazole susceptibility is a strong marker of isavuconazole susceptibility in most clinically relevant Candida and Aspergillus isolates. Of note, this suggests that the azole agar screening method (EUCAST E.Def 10.1) can be adopted for identification of A. fumigatus isolates suitable as targets for isavuconazole therapy despite the fact that an isavuconazole agar is not included in the plate design (13).
Isavuconazole is licensed as a second-line option for the treatment of Mucorales infections in adults after the VITAL study showed equal clinical efficacy compared to that for matched historical controls treated with amphotericin B (14). In that study, species-specific outcome evaluation was not performed, probably in part because one-third of the cases lacked species identification. We found consistently high MICs of Ն16 mg/liter for M. circinelloides, confirming previous findings by CLSI and EUCAST testing (15,16). Hence, species identification is highly recommended, as clinical efficacy remains to be confirmed for this species.
Isavuconazole is not licensed for the treatment of invasive candidiasis after a recent phase 3, randomized, double-blind, multinational clinical trial failed to demonstrate noninferiority at the end of intravenous (i.v.) therapy compared to caspofungin (17). Thus, the trial supported the results of other clinical studies showing superiority of echinocandins over azoles and amphotericin B and, thus, the recommendation of echinocandins as first-line agents for candidemia and invasive candidiasis (18)(19)(20)(21). Nevertheless, the secondary endpoints (overall response to therapy 2 weeks after the end of therapy and all-cause mortality on days 14 and 56) were similar between arms, as were safety and median time to clearance from the bloodstream. On this background and taking the potent in vitro activity and attractive safety profile compared to fluconazole and voriconazole into account, isavuconazole might serve as a valid second-line option in settings where echinocandin resistance is likely or documented, mold coverage is indicated, or oral therapy is preferred.
In summary, isavuconazole displayed broad in vitro activity against most humanpathogenic species, including dermatophytes and several uncommon species. Of note, however, we confirmed low isavuconazole in vitro activity against M. circinelloides and therefore advocate for performing species identification, also for Mucorales, whenever possible. Acquired isavuconazole resistance was infrequent, except in A. terreus, C. glabrata, and C. tropicalis, and, when present, was associated with cross-resistance to other azoles. Continued surveillance remains important.

MATERIALS AND METHODS
Isolates. In total, 1,069 yeast and 429 mold isolates from 1,325 patients were included (1 isolate each from 1,186 patients and 2 to 7 isolates from 139 patients). The isolates were prospectively obtained from clinical samples or pure cultures received at the mycology reference laboratory at Statens Serum Institut for identification and susceptibility testing during 2016 and 2017. No ethical restraints apply to studies of routinely obtained anonymized laboratory data. Same-species isolates from the same patient were excluded from the study if sampled Յ21 days apart and identical MICs (within Ϯ1 dilution step) were seen. The isolates derived from the entire country and the following clinical specimens: blood as part of the national surveillance program (915 specimens), airways/lung/pleura/sinus (404), other normally sterile sites (47), urine (20), skin/scalp/nail (24), cervix/vagina/urethra (13), other superficial sites (39), and other/unspecified (36). Yeast identification was done using macro-and micromorphology, supplemented by thermotolerance (incubation at 37°C and 43°C), matrix-assisted laser desorption ionization-time of flight mass spectrometry (Bruker, Bremen, Germany) for Candida (22), and, when needed, internal transcribed spacer (ITS) sequencing (23). Similarly, mold identification was done by classical techniques, including thermotolerance (incubation at 50°C) for discriminating A. fumigatus sensu stricto from cryptic species, which underwent ␤-tubulin sequencing (24). The use of the term "complex" is acknowledged for Aspergillus species other than A. fumigatus, in the absence of detailed molecular identification, although for simplicity, it is not used throughout this work. ITS and TEF (transcription elongation factor) sequencing were adopted for other molds and Fusarium species, specifically (23,25).
Data analysis. Modal MICs, MIC 50 s, geometric mean MICs (GM-MICs), and MIC ranges were determined for individual species (n Ն 10). EUCAST ECOFFs/breakpoints were adopted for wild-type/susceptibility classification. For species without EUCAST ECOFFs, MICs Ͼ2 dilution steps above the modal MIC were regarded as non-wild type. However, for species where the modal MIC was equal to or lower than the lowest concentration tested, MICs Ն2 dilution steps above the modal MIC were regarded as non-wild type. Pearson correlation analyses with a two-tailed P value were performed for comparisons of antifungal in vitro activities (after log 2 transformation) using GraphPad Prism 7.04. Correlation coefficients (squared) (R 2 ) of Ն0.5 with a P value Ͻ0.05 were interpreted as a significant indicator of good to strong correlation, whereas R 2 values of Ͻ0.5 indicated weak correlation.