Susceptibility to echinocandins of Candida spp. strains isolated in Italy assessed by European Committee for Antimicrobial Susceptibility Testing and Clinical Laboratory Standards Institute broth microdilution methods

Background The echinocandins are recommended as first-line therapy for Candida species infections, but drug resistance, especially among Candida glabrata, is becoming more frequent. We investigated the antifungal susceptibility of anidulafungin, caspofungin, and micafungin against 584 isolates of Candida spp. (bloodstream, other sterile sites) collected from patients admitted to an Italian university hospital between 2000 and 2013. The susceptibility was evaluated using the broth microdilution method according to both the European Committee for Antimicrobial Susceptibility Testing (EUCAST EDef 7.2) and the Clinical Laboratory Standards Institute (CLSI M27-A3). The echinocandin susceptibilities were assessed on the basis of the species-specific clinical breakpoints proposed by the EUCAST version 6.1 and CLSI M27-S4 documents. The two methods were comparable by assessing essential agreement (EA), categorical agreement (CA), and Spearman’s correlation analysis (rho, r). Results The modal minimum inhibitory concentrations (MICs; μg ⋅ mL −1) values by both methods (EUCAST/CLSI) for anidulafungin, caspofungin, and micafungin for each species were, respectively, as follows: C. albicans, 0.03/0.12, 0.016/0.5, and 0.016/0.008; C. parapsilosis complex, 2/1, 2/2, and 2/1; C. tropicalis, 0.06/0.12, 0.06/0.12, and 0.06/0.12; C. glabrata complex, 0.03/0.25, 0.06/0.12, and 0.03/0.06; C. guilliermondii, 2/1, 2/2, and 2/2; and C. krusei, 0.06/0.12, 0.12/0.5, and 0.06/0.12. The overall resistance rates for EUCAST/CLSI were as follows: anidulafungin, 2.5/0.9 %; caspofungin, breakpoint not available/3.8 %; micafungin, 2.7/1.5 %. Candida glabrata complex was the least susceptible to all three echinocandins, and the percentages of resistant isolates by EUCAST/CLSI were as follows: anidulafungin, 13.5/2.7 %; caspofungin, breakpoint not available/16.2 %; micafungin, 18.9/13.5 %. The overall EA was 93 % for micafungin, 92 % for anidulafungin, and 90 % for caspofungin. The CA was >90 % for all organism-drug combinations with the exception of C. glabrata and anidulafungin (89 %). Spearman’s rho for EUCAST/CLSI was 0.89 (p < 0.001) for caspofungin, 0.85 (p < 0.001) for anidulafungin, and 0.83 for micafungin (p < 0.001). Conclusions Independent of the procedure applied, no alarming resistance to the tested agents was found, although a reduced susceptibility was detected for C. glabrata complex. The EUCAST and CLSI methods produce similar MICs, indicating that using one method or the other should not result in susceptibilities different enough to affect treatment decisions.


Background
The echinocandins (anidulafungin [AND], caspofungin [CSP], and micafungin [MCF]) are lipopeptides that inhibit glucan synthase, which is responsible for the biosynthesis of β-1,3-D-glucan, a major structural component of fungal cell walls. These drugs demonstrate fungicidal activity against most species of Candida and are effective against azole-resistant yeasts and Candida-forming biofilms [1][2][3][4][5]. The Infectious Diseases Society of America [6] and the European Society of Clinical Microbiology and Infectious Diseases [7] guidelines for the management of Candida infection recommend echinocandins for firstline therapy [8].
Until now, in Italy, there is not an extensive study concerning the in vitro susceptibility of echinocandins against Candida isolates using the EUCAST method [14]. The aims of this study are: i) to determine the susceptibilities of Candida spp. to AND, CSP, and MCF using both the EUCAST and CLSI methods; and ii) to compare the performance of both methods by assessing the essential agreement (EA) and categorical agreement (CA) levels.

Results and discussion
In each experiment, the MIC values of the quality control strains fell within the established ranges published for both methods [9,13]. The modal MIC (μg ⋅ mL − 1 ; 15 repetitions) values by EUCAST/CLSI for AND, CSP, and MCF for each control strain were, respectively, as follows: C. krusei ATCC 6258, 0.06/0.12, 0.5/0.5, and 0.12/0.12; C. parapsilosis ATCC 22019, 1/1, 0.5/1, and 0.5/0.5. Table 1 summarizes the in vitro susceptibility values of 584 clinical isolates of Candida species to echinocandins as determined by the CLSI and EUCAST BMD methods, as well as EA, and CA between two methods for each echinocandin and Candida species. Generally, the MIC values for all three echinocandins were low and below the susceptibility breakpoint, regardless of the method used. As shown by other authors [15], the MIC values for AND (geometric mean MIC EUCAST/CLSI, 0.16/0.22 μg ⋅ mL −1 ) and MCF (geometric mean MIC EUCAST/CLSI, 0.13/0.14 μg ⋅ mL −1 ) were lower those for CSP (geometric mean MIC EUCAST/CLSI, 0.29/0.33 μg ⋅ mL −1 ) by both assay, suggesting that they have superior in vitro potency. When the species-specific clinical breakpoints were applied, the overall resistance rates for EUCAST/CLSI were as follows: AND, 2.5/0.9 %; CSP, breakpoint not available/ 3.8 %; MCF, 2.7/1.5 %. These data are consistent with those reported previously [14,[16][17][18] and document the excellent potency and spectrum of echinocandins against most Candida spp. Of the 22 CSP resistant isolates by CLSI, five were MCF resistant, three were AND resistant, and one was resistant both AND and MCF; this discrepant susceptibility pattern is according to other studies [17,19]. Given the mechanism of action that is shared among the echinocandins, it is biologically unlikely that such large percentages of isolates are non susceptible to CSP but remain susceptible to AND and MCF. A potential explanation for this finding may be the technical issues associated with the in vitro testing of CSP rather than a true difference in antifungal activity [20]. For this reason, neither EUCAST nor CLSI procedures recommend the use of CSP for antifungal susceptibility testing, while AND and MCF as markers for CSP susceptibility [11,21].
The trend in the rate of resistance to echinocandins was analyzed: no significant trend was observed for each drug over the studied 14-year period, by both methods.
The MIC values were the highest for C. parapsilosis complex and C. guilliermondii by the two methods. A natural polymorphism occurring in the hot spot, one region of FKS1 has been suggested to be responsible for the reduced echinocandin susceptibilities of these species [22]. Unlike yeasts with acquired FKS mutations, these strains respond well to standard therapy presumably because the polymorphism only weakly affects the sensitivity of glucan synthase for drug [5]. Moreover, the good clinical response of C. parapsilosis may be due to its less virulent, reduced capacity to invade the deep tissue, and the high probability of therapeutic success if central venous catheter is removed [15].
For most isolates, the echinocandin MICs obtained by the EUCAST method tended to be one twofold dilution lower than those obtained by the CLSI method. The reason for this result could be related to known differences between the two methods: the higher carbohydrate content in the RPMI and higher inoculum could be responsible for the lowering of the EUCAST results [28].
The overall EA was very high: 93 % for MCF, 92 % for AND, and 90 % for CSP. The Spearman's correlation analysis shows a significant positive correlation between EUCAST and CLSI MICs (r = 0.85, p < 0.001 for AND; r = 0.89, p < 0.001 for CSP; r = 0.8, p < 0.001 for MCF). This finding was consistent with the results of previous global multicenter studies by Pfaller et al. [29,30] and confirms the high level of EA between the reference procedures. The rates of EA were also high when results were analyzed per species, and the worst EA (72 %) was for C. glabrata complex tested against AND. A good CA was also observed for all organism-drug combinations, ranging from 89 to 100 %. The lowest CA was for C. glabrata complex isolates to AND, where five (13.5 %) isolates were resistant when applying EUCAST breakpoints as opposed to one (2.7 %) when applying the CLSI breakpoints. The meaning of this in vitro finding is not clear and needs to be clarified in more detail. The best CA was for C. parapsilosis complex to AND and MCF.

Conclusions
Our study had some limitations. First, none of our strains were characterized with respect to echinocandin resistance mechanisms. Second, the lack of EUCAST species-specific breakpoints for CSP precludes a more standard comparison for the assessment of the CA. Third, this study was an observational laboratory based survey, therefore no data was available as to the type and duration of antifungal therapy. Nevertheless, to the best of our knowledge, this is the first study on this topic to be carried out in Italy. Our data show that the EUCAST and CLSI echinocandin MIC values for Candida isolates from patients in Italy are in accordance with the worldwide epidemiology of Candida strains [14,[16][17][18]. Likewise, the echinocandins were active in vitro against the majority of Candida species tested in this study; a reduced susceptibility, as has been noted elsewhere [23,24], was detected for C. glabrata complex and C. tropicalis. MIC values obtained by the CLSI and EUCAST methods are comparable for the testing of echinocandins against Candida species. Regression analysis confirms the close proximity of the MICs generated by each method. In most cases, the MIC differences between standard procedures are small enough that using one method or the other should not result in susceptibilities that are different enough to affect treatment decisions. The clinical implications regarding the improvement of the susceptibility tests to echinocandins are significant, since accurate data are important in defining differences in clinical efficacy among AND, CSP and MCF, thereby supporting the most appropriate choice of early antifungal treatment towards a better prognosis. Further efforts are needed to harmonize the two standard procedures.

Clinical isolates
Between January 2000 and December 2013, a total of 597 clinical isolates of Candida spp. (bloodstream and other sterile sites) were collected from patients admitted to a large Italian university hospital. Of these, the six most common Candida species were tested against echinocandins (251 C. albicans, 224 C. parapsilosis, 46 C. tropicalis, 37 C. glabrata, 15 C. guilliermondii, 11 C. krusei), for a total of 584 isolates recovered from the following wards: intensive care unit (n = 288), haematology (n = 99), internal medicine (n = 88), surgery (n = 85), and oncology (n = 24). For this study, we did not use any additional data or samples other than those obtained through routine laboratory collection. Therefore, neither ethical approval nor patient consent was considered necessary. Registered data were managed in accordance with the Italian data protection laws (privacy law). The isolates were identified using standard procedures (i.e., morphology on cornmeal agar plates, germ-tube production in serum, and ability to grow at 37°C and 42°C) and biochemical analysis using the ID32C and VITEK-2 System (Biomérieux, Marcy l'Etoile, France). Each isolate represented a unique strain from a single patient and was frozen at −80°C until the analysis. Prior to being tested, each isolate was subcultured on Sabouraud dextrose agar plates (BioMèrieux) to ensure purity, viability, and optimal growth characteristics.

Susceptibility testing
AND (Pfizer Pharmaceuticals, Groton, CT, USA), CSP (Merck & Co., Inc., Whitehouse Station, NJ, USA), and MCF (Astellas Pharma, Tokyo, Japan) were obtained as standard powders. A single lot of pure substance for each of the three echinocandins was used. Stock solutions were prepared in dimethyl sulfoxide (Sigma, St. Louis, MO, USA) for each echinocandin, taking into account the potencies of the powders.
CLSI BMD testing was performed according to document M27-A3 [9,10] using RPMI 1640 medium with 0.2 % glucose (Sigma) and 0.165 M MOPS, an inoculum of 0.5 × 10 3 to 2.5 × 10 3 cells ⋅ mL −1 , and incubation at 35°C. The MIC values were determined visually after 24 h of incubation as the lowest concentration of drug that caused a significant diminution (≥50 % inhibition) of growth below control levels.
EUCAST BMD testing was performed according to document EDef 7.2 [available on the EUCAST website: http://www.eucast.org] using RPMI 1640 medium supplemented with glucose to a final concentration of 2 %, an inoculum of 0.5 × 10 5 to 2.5 × 10 5 cells ⋅ mL − 1 , and incubation at 35°C. MIC values were determined with a spectrophotometer (wavelength of 450 nm; ETI System Fast Reader ELX, Biotek, US) after 24 h of incubation as the lowest concentration of drug that resulted in >50 % inhibition of growth relative to that of the growth control.
C. krusei ATCC 6258 and C. parapsilosis ATCC 22019 were used as quality control strains in each run according to the CLSI M27-A3 document [9].

Interpretation and analysis of results
The echinocandin susceptibilities were defined according to the species-specific clinical breakpoints proposed by the EUCAST version 6.1 [31] and CLSI M27-S4 documents [10] (Table 2). Differences in the temporal susceptibility rates were analysed by chi-square test for trend. MIC discrepancies of no more than ± 2-fold dilutions were used to calculate the EA. The CA was defined as the percentage of the discrepancy in the number of resistant isolates based on the existence of interpretative breakpoints [16]. Moreover, to determine the correlation between the methods, Spearman's rank correlation coefficient (rho, r), and its corresponding p value was performed by plotting EUCAST versus CLSI MICs. The level of significance was set at a p value less than 0.05. Statistical analysis of data was carried out using STATA MP 11.2 for Mac Os X (SPSS Inc., Chicago, IL, USA).