Previous Article | Next Article ![]()
Journal of Clinical Microbiology, August 2005, p. 3884-3889, Vol. 43, No. 8
0095-1137/05/$08.00+0 doi:10.1128/JCM.43.8.3884-3889.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
VCU Medical Center, Richmond, Virginia,1 Public Health, Ancona, Italy,2 Instituto de Salud Carlos III, Majadahonda, Spain,3 University of Iowa College of Medicine, Iowa City, Iowa,4 University of Texas Health Science Center, San Antonio, Texas,5 Medical Center Nijmegen, Nijmegen, The Netherlands6
Received 14 February 2005/ Returned for modification 11 April 2005/ Accepted 20 May 2005
|
|
|---|
92%. When the comparison was based on CLSI breakpoint categorization, the agreement was 68 to 76% for three of the four species that included fluconazole-resistant and S-DD isolates; 9% very major discrepancies (
8 µg/ml versus
64 µg/ml) were observed among fluconazole-resistant isolates and 50% with voriconazole (
1 µg/ml versus
4 µg/ml). Similar results were observed with itraconazole for seven of the eight species evaluated (28 to 77% categorical agreement). Posaconazole EUCAST MICs were also substantially lower than CLSI MIC modes (0.008 to 1 µg/ml versus 1 to
8 µg/ml) for some of these isolates. Therefore, the CLSI breakpoints should not be used to interpret EUCAST MIC data. |
|
|---|
The purposes of the present study were (i) to compare EUCAST and reference CLSI MICs of fluconazole, itraconazole, posaconazole, and voriconazole for 71 Candida isolates; (ii) to assess the reproducibility among six laboratories of MIC results obtained by each method; (iii) to determine the utility of CLSI fluconazole, itraconazole, and voriconazole breakpoints for EUCAST MIC data; and (iv) to identify substantial differences (>3 dilutions) between the methods for posaconazole MICs. Five to 15 isolates of Candida albicans, C. dubliniensis, C. glabrata, C. guilliermondii, C. krusei, C. lusitaniae, C. parapsilosis, and C. tropicalis were evaluated in each center.
|
|
|---|
|
View this table: [in a new window] |
TABLE 1. Agreement between EUCAST and CLSI MICs for Candida spp. in six laboratoriesa
|
|
View this table: [in a new window] |
TABLE 2. Agreement between EUCAST and CLSI fluconazole and itraconazole MICs according to CLSI breakpoint categorization for Candida spp. in six laboratoriesa
|
|
View this table: [in a new window] |
TABLE 3. Substantially discrepant posaconazole EUCAST MICs compared to CLSI resultsa
|
Antifungal agents. EUCAST and CLSI reference microdilution plates containing serial drug dilutions of posaconazole (Schering-Plough Research Institute, Kenilworth, N.J.), voriconazole (Pfizer Central Research, New York, N.Y.), and the established triazoles fluconazole (Pfizer Central Research) and itraconazole (Janssen, Beerse, Belgium) were prepared by TREK Diagnostic Systems (Cleveland, OH) following the CLSI M27-A2 and EUCAST guidelines (4, 6), shipped frozen to each laboratory, and stored at 70°C until the day of the test. Voriconazole and itraconazole drug dilutions ranged from 0.008 to 16 µg/ml, posaconazole from 0.004 to 8 µg/ml, and fluconazole from 0.12 to 128 µg/ml in both reference and EUCAST microdilution plates.
Stock inoculum preparation. Stock inoculum suspensions of the yeasts were prepared in sterile saline (8.5 g/liter) NaCl from 24-h cultures on Sabouraud dextrose agar at 35°C. The turbidity of each yeast suspension was adjusted by the spectrophotometric method (6).
CLSI broth microdilution method (M27-A2 document). U-bottom microdilution plates containing 100 µl of the twofold serial dilutions of the antifungal drugs in standard RPMI 1640 medium (0.2% glucose) were inoculated with 100 µl of inoculum containing between 1.0 x 103 and 5 x 103 CFU/ml. Following inoculation of the reference microdilution plates, they were incubated at 35°C in a non-CO2 incubator, and MICs were determined after 24 and 48 h. Reference MICs corresponded to the lowest drug dilution that showed prominent growth inhibition (50% or more) (6). QC isolates were tested in the same manner in each participant laboratory.
Proposed EUCAST broth microdilution method.
Flat-bottom microdilution plates containing 100 µl of the twofold serial dilutions of the antifungal drugs in double-strength RPMI 1640 medium (2% glucose) were inoculated with 100 µl of inoculum containing between 0.5 x 105 and 2.5 x 105 CFU/ml. The microdilution plates were incubated at 35°C in a non-CO2 incubator, and MICs were determined after 24 h; microdilution plates were reincubated if the optical density was
0.5 (indicative of poor growth) and read after the second day of incubation. MICs were determined with a spectrophotometer at a wavelength in the range of 530 to 550 nm. EUCAST MICs corresponded to the lowest drug dilution that showed a reduction of growth of 50% or more compared with the growth control (4). QC isolates were tested in the same manner in each participant laboratory.
Statistical analysis. Both on-scale (e.g., 0.12 and 128 µg/ml) and off-scale (i.e., <0.12 and >128 µg/ml) MICs were included in the analysis. For the comparison between the two methods, MICs of each drug-organism combination obtained by each method in the six laboratories were compared as follows: (i) 24-h MIC pairs by both methods and (ii) EUCAST MICs versus CLSI 48-h MICs. Values were considered in agreement when the discrepancies between the methods were no more than 2 log2 dilutions. The reproducibility of the results obtained by the six laboratories was evaluated by determining the percent agreement between MICs that were within 3 dilutions (e.g., 0.25, 0.5, and 1 µg/ml) as well by intraclass correlation coefficients (ICCs) for converted log2 MICs. In addition, because the EUCAST Subcommittee has not yet established breakpoints, the interpretive CLSI criteria (M27-A2 document) were used to evaluate the agreement between EUCAST and CLSI results for fluconazole and itraconazole data regarding these interpretative criteria (6). Tentative interpretive breakpoints have recently been established for voriconazole (personal communication, CLSI Subcommittee, January 2005 meeting) but are not available for posaconazole. Therefore, substantial differences (4 or more dilutions) between the methods were also identified for the latter agent.
|
|
|---|
Table 1 lists the percentages of agreement between the two methods for the eight Candida spp. evaluated as well as the percentages of interlaboratory reproducibility for each method. These percentages represent agreement in MIC results of no more than 2 log2 dilutions, regardless of breakpoint classification. The agreement between the methods was usually higher when testing fluconazole and voriconazole (92 to 95% overall agreement; ICCs, 0.73 to 0.95) than when testing the other two triazoles (85 to 91% overall agreement; ICCs, 0.56 to 0.93). The compatibility of EUCAST with CLSI MICs was slightly higher with 24-h results than with 48-h results (90 to 95% versus 85 to 94%, respectively). The lowest percentages of agreement were observed for isolates of C. albicans, C. glabrata, and C. tropicalis. The interlaboratory agreements among the six laboratories of results obtained by each method were similar (92 to 95% EUCAST versus 93 to 97% CLSI); the lowest percentages of interlaboratory reproducibility of EUCAST MICs were observed for isolates of C. glabrata and C. krusei (86 to 88%) when testing itraconazole and posaconazole and of the CLSI method when testing itraconazole and voriconazole against C. albicans and C. glabrata (81 and 86%, respectively).
Table 2 depicts the percentages of agreement between CLSI 48-h and EUCAST MICs regarding the categorical data established by the CLSI for fluconazole and itraconazole (6). Low rates of agreement were shown for four of the species that included either resistant or S-DD isolates (68 to 90% agreement with fluconazole and 28 to 77% with itraconazole), because EUCAST MICs were substantially and consistently lower than CLSI results. The agreement was good (96 to 100%) for C. dubliniensis, C. lusitaniae, C. parapsilosis, and C. tropicalis. In the majority of the cases, the EUCAST method categorized an isolate as either susceptible or S-DD, while by the CLSI method the isolate belonged to either the S-DD or resistant categories, respectively. However, major discrepancies (9%) were also observed where five fluconazole EUCAST susceptible values were obtained among 58 CLSI resistant MICs. Similar results were obtained with itraconazole, but the percentages of agreement were lower (32 to 73%) for most of the species.
With voriconazole, 18 susceptible MICs (
1 µg/ml) were obtained by the EUCAST method among 36 resistant values (
4 µg/ml) by the reference method. Substantially lower EUCAST than CLSI MICs were also documented for three isolates of C. albicans, C. glabrata, and C. tropicalis for which CLSI posaconazole mode MICs were >1 µg/ml (Table 3). The majority (90%) of posaconazole MICs were
1 µg/ml for 5 of the 10 fluconazole-resistant isolates compared to 9% by the EUCAST method.
Table 4 depicts the susceptibilities of the 71 isolates to the four agents in the six laboratories by both methods. In general, MICs by both methods appear to be similar, as demonstrated by the wide MIC range for certain species and drug combinations. However, in some instances for the species that included fluconazole-resistant isolates, fluconazole and itraconazole MICs at which 90% of the isolates tested were inhibited (MIC90s) were in the S-DD category, while the CLSI corresponding results were in the resistant category. Although posaconazole and voriconazole MIC90s were within 1 dilution between the two methods for most of the species, CLSI MIC90s for C. albicans, C. glabrata, and C. guilliermondii tended to be
2 µg/ml, while EUCAST MICs were
1 µg/ml.
|
View this table: [in a new window] |
TABLE 4. Fluconazole, itraconazole, posaconazole, and voriconazole EUCAST and CLSI MIC data for Candida isolatesa
|
|
|
|---|
The agreement between the methods was dependent on the species, the antifungal agent, and the incubation time used to determine the CLSI results. The highest agreements were observed with fluconazole and voriconazole. Three previous studies have evaluated the compatibility between EUCAST and CLSI microdilution methods in one to two laboratories (1-3). Although similarly good reproducibility (overall, 92 to 99%) has been reported with fluconazole, itraconazole, and voriconazole (1, 2), in another study (3), the reproducibility was lower with itraconazole (78 to 81%) and for C. tropicalis with fluconazole (88%). In our study, itraconazole MICs also yielded the lowest agreement between the methods (Table 1). Results with posaconazole were adequate but had consistently lower agreement (79 to 100%) than that obtained with fluconazole and voriconazole, which is not surprising, since this antifungal is more closely related to itraconazole than to fluconazole. Comparisons of the two methods with posaconazole have not been reported. Among the species, the lowest percentages of reproducibility between the methods were observed for C. albicans, C. krusei, and C. tropicalis with either itraconazole, posaconazole, or both. Cuenca-Estrella et al. (3) also found low (88%) reproducibility between the methods for C. tropicalis versus fluconazole.
The interlaboratory reproducibility of each method among the six laboratories was mostly good to excellent (Table 1). Similar results were obtained among nine laboratories (4) with fluconazole for six of eight isolates of Candida spp. (98 to 100%) and for all isolates with itraconazole (85 to 93%). When testing fluconazole, the lowest interlaboratory agreement of EUCAST MICs (83%) was for the C. glabrata isolate (4), while in our study, the agreement was 95% for this species.
Antimicrobial susceptibility testing should not only provide reproducible data butalso identify isolates that are potentially resistant to the agent being evaluated. Because of that, we included 10 strains, among the different species, that have been classified by the CLSI method as resistant to fluconazole (MICs,
64 µg/ml). These fluconazole results were duplicated by the six laboratories by the CLSI method, but three EUCAST results were
8 µg/ml (susceptible values); the same applied for isolates that were categorized as itraconazole resistant by the CLSI method. Similar categorical discrepancies between these two methods have been reported by Cuenca-Estrella et al. (3) with both fluconazole (C. parapsilosis and C. glabrata) and itraconazole (C. glabrata, C. krusei, and C. tropicalis). Because in our study major categorical discrepancies were not observed with 24-h CLSI MICs, the higher carbohydrate content in the RPMI 1640 could be responsible for the lowering of EUCAST results. Since EUCAST MICs were consistently the lower values and trays were not shaken prior to the spectrophotometric reading, it is possible that the unusually elevated EUCAST results (0.5%) were caused by the presence of air bubbles in the MIC wells, as has been demonstrated in our laboratory (personal communication).
Interpretative breakpoints are not available for posaconazole, but tentative MIC breakpoints of
1 µg/ml (susceptible) and
4 µg/ml (resistant) have been recently established for voriconazole (personal communication, CLSI Subcommittee, January 2005 meeting). In our study, 50% of voriconazole resistant MICs corresponded to susceptible values by the EUCAST method; these could be considered very major errors. EUCAST MICs substantially lower than CLSI MICs have been previously reported for C. glabrata with voriconazole (MIC90s of 0.5 and 4 µg/ml, respectively) and for C. parapsilosis with caspofungin (MIC ranges, 0.5 to 2 and 0.5 to >16 µg/ml, respectively) (1). Although the two methods have not been compared for testing posaconazole, a similar trend was observed with this agent for some isolates (Table 4).
In conclusion, results of this study and previous comparisons of both methods indicate that EUCAST MICs of the triazoles are consistently lower than those obtained with the CLSI reference method. Because these lower EUCAST values also reflect substantial categorical shifting, including major discrepant results, the CLSI breakpoints do not appear to be useful for the categorical interpretation of EUCAST MIC data. Therefore, it is not recommended to provide interpretive MIC results when reporting EUCAST in vitro data until breakpoints have been established for EUCAST methodology.
We thank Mercedes Ramirez and Ana L. Rodriguez for their technical assistance.
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»