|Home | About | Journals | Submit | Contact Us | Français|
Anidulafungin Etest and CLSI MICs were compared for 143 Candida sp. isolates to assess essential (within 2 log2 dilutions) and categorical agreements (according to three susceptibility breakpoints). Based on agreement percentages, our data indicated that Etest is not suitable to test anidulafungin against Candida parapsilosis and C. guilliermondii (54.4 to 82.4% essential and categorical agreements) but is more suitable for C. albicans, C. glabrata, C. krusei, and C. tropicalis (87.9 to 100% categorical agreement).
The echinocandins are available for intravenous treatment of Candida infections, especially for patients with recent azole exposure (17, 26, 27). The Clinical and Laboratory Standards Institute (CLSI) has established guidelines and an interpretive susceptibility breakpoint (≤2 μg/ml) for testing echinocandins against Candida spp. (11, 12). We evaluated the suitability (essential and categorical agreements) of anidulafungin Etest MICs for 143 Candida sp. bloodstream isolates from the Hospital La Fe, Valencia, Spain. Categorical agreement was evaluated according to CLSI (11), Garcia-Effron et al. (21), and Desnos-Ollivier et al. (13, 14) susceptibility microdilution breakpoints (≤2, ≤0.5, and ≤0.25 μg/ml, respectively).
The 143 isolates included caspofungin-resistant (six heterozygous and homozygous C. albicans mutants, one C. krusei isolate), caspofungin-susceptible (Table (Table1),1), and quality control (QC; C. parapsilosis ATCC 22019 and C. krusei ATCC 6258) isolates (8, 15, 23); anidulafungin MICs were within the established QC limits (12).
Anidulafungin (Pfizer, Madrid, Spain) MICs were determined for the 143 isolates (Table (Table2)2) by both the CLSI M27-A3 and Etest methods after 24 h at 35°C (11). Reference microdilution trays containing serial drug dilutions (0.016 to 8 μg/ml) in RPMI 1640 medium (0.2% glucose; Sigma-Aldrich, Madrid, Spain) were inoculated with a 1 × 103- to 5 × 103-CFU/ml inoculum. MICs were the lowest drug dilutions that showed ≥50% inhibition (11). Etest MICs were determined according to the manufacturer's instructions (AB BIODISK, Solna, Sweden) using RPMI agar (2% glucose), an approximately 1 × 106- to 5 × 106-CFU/ml inoculum, and Etest strips (0.002 to 32 μg/ml). MICs were the lowest drug concentrations at which the border of the elliptical inhibition intercepted the strip scale, ignoring trailing growth.
Anidulafungin MICs were in essential agreement when the discrepancies between the two methods were within 2 dilutions. Categorical errors were calculated according to each of the three breakpoints as follows: (i) very major errors when the reference MIC indicated resistance while Etest indicated susceptibility and (ii) major errors when the Etest categorized the isolate as resistant and the reference as susceptible. For the correlation between the methods, a linear regression analysis using the least-squares method (Pearson's correlation coefficient; MS Excel software) was performed by plotting Etest versus reference MICs.
Echinocandin resistance in Candida spp. has been associated with high MICs, mutations in the FKS1 gene, and therapeutic failure (2, 4, 8, 13, 14, 20, 25). MICs higher than those for other species are consistently observed for C. parapsilosis and C. guilliermondii (28), along with reduced glucan synthase sensitivity (19) and a lack of killing activity for C. guilliermondii (5, 6, 7). Based on both reproducibility and the ability to discriminate between wild strains and caspofungin-resistant mutants (Table (Table1),1), the CLSI established standard conditions for testing echinocandins against Candida spp. (11, 12); we followed this methodology. The evaluation of a new assay requires both essential and categorical agreements; the latter was accomplished using CLSI (11) and two other nonsusceptible microdilution breakpoints (>2, >0.5, and ≥0.5 μg/ml) (13, 14, 21).
Our CLSI MIC data for most species were similar to those previously published (28), as demonstrated by our MIC90s (MICs for 90% of the isolates tested), except for C. albicans. However, most of the Etest MIC90s were higher than the CLSI results in this and another study (28), which impacted both agreements (Table (Table2).2). Although the overall essential agreement was 79.7% (R, 0.82; Fig. Fig.1),1), it was >90% for two of the six species (Table (Table2)2) and similar to prior Etest and CLSI comparisons for caspofungin and C. albicans (91 versus 89%), higher for C. tropicalis (100 versus 88%), and lower for the other four species (69.2 to 77.8% versus 90 to 100%) (1, 30). Caspofungin Etest MICs usually were lower than the reference results for yeasts (1, 9, 30) and Aspergillus spp. (16), but our anidulafungin Etest MICs were mostly higher. Although lack of prior evaluations precluded comparisons, the acceptable essential agreement is ≥90% (10). It is unfortunate that C. parapsilosis and C. guilliermondii were among the species with unsuitably low essential agreement, because little information has been gathered in either efficacy clinical trials or molecular studies (5, 22, 24, 31). High MICs (>0.5 μg/ml) were not observed for the clinical isolates of the other species where the essential agreement was low, and therefore those results did not affect the categorical agreement (Tables (Tables22 and and33).
The categorical agreement was suitable (87.9 to 100%) for four of the six species evaluated, breakpoint dependent (11, 13, 14, 21) (Table (Table3).3). Again, the lowest percentages were for C. parapsilosis and C. guilliermondii (54.4 to 82.4% according to breakpoint) due to major errors (8.8 to 36.8% false resistance) and very major errors (8.8 to 14% false susceptibility for C. parapsilosis only). The best categorical agreement for C. albicans was according to the CLSI breakpoint (11). The FDA target for major errors is ≤3% and ≤1.5% for very major errors (18). Therefore, Etest could be considered unsuitable for testing of C. parapsilosis and C. guilliermondii with anidulafungin but suitable for the other four species (Table (Table3).3). Categorical agreement was not assessed during prior Etest caspofungin evaluations (1, 9, 30), but the agreement was >99% for echinocandin YeastOne MICs for Candida spp. (29).
Etest has detected echinocandin resistance (fks1 gene mutations) among Candida and Aspergillus species (2, 3, 4, 14), but similar MICs were obtained by reference methodology for Candida spp. (2, 4). While these results confirmed the lower susceptibility breakpoint (≤0.5 μg/ml) for micafungin and anidulafungin versus C. albicans (21), it is uncertain if this endpoint is applicable for either C. parapsilosis or C. guilliermondii. The CLSI susceptibility breakpoint (≤2 μg/ml) was based on clinical trial data, global susceptibility surveillance, resistance mechanisms, and pharmacokinetic and pharmacodynamic parameters from model systems (11, 28). The response to therapy has been comparable for Candida species, but few isolates of C. parapsilosis (9 to 10%) and C. guilliermondii (none) were included in anidulafungin clinical trials (24, 31). More information is needed for these species; the response of most C. parapsilosis infections to echinocandin therapy, regardless of the reduced susceptibility of these two species, could be due to their lower virulence.
In conclusion, our preliminary data indicated unsuitable percentages of both essential and categorical agreements for C. parapsilosis and C. guilliermondii. To our knowledge, Etest has not been evaluated in multicenter studies to assess its reliability and ability to identify echinocandin resistance. Such studies with large numbers of isolates, including well-documented resistant isolates, are essential before using Etest routinely.
Etest strips were provided by Pfizer, Spain, and this study was also partially supported by a Pfizer, Spain, research grant (VOR.IIG-48).
Published ahead of print on 22 December 2009.