Multicenter Comparison of the Sensititre YeastOne Colorimetric Antifungal Panel with the NCCLS M27-A2 Reference Method for Testing New Antifungal Agents against Clinical Isolates of Candida spp.

A multicenter (three centers) study compared MICs obtained bythe Sensititre YeastOne Colorimetric Antifungal plate to referencemicrodilution broth (NCCLS M27-A2 document) MICs of three newtriazoles (posaconazole, ravuconazole, and voriconazole) andthe echinocandin caspofungin acetate for 100 isolates of Candidaspp. In addition, amphotericin B and fluconazole were testedas control drugs. Colorimetric MICs of caspofungin and amphotericinB corresponded to the first blue well (no growth), and MICsof the other agents corresponded to the first slightly purpleor blue well. Two comparisons of MIC pairs by the two methodswere evaluated: 24-h colorimetric MICs were compared to NCCLSMICs at 24 and at 48 h. The interlaboratory reproducibilityof YeastOne and reference MICs was also examined. The best performanceof the YeastOne plate was with 24-h MICs (overall, 95 to 99%agreement) for all the species and antifungal agents. Theseresults suggest the potential value of the YeastOne plate foruse in the clinical laboratory for the four new antifungal agentsevaluated.

INTRODUCTION

Top

Introduction

The National Committee for Clinical Laboratory Standards hasdeveloped a reference microdilution broth method (NCCLS documentM27-A2) for antifungal susceptibility testing of yeasts (9).Alternatively, more simplified and efficient approaches alsohave been developed. Among them, the Sensititre YeastOne ColorimetricAntifungal plate (TREK Diagnostics Systems, Cleveland, Ohio)is a disposable plate that contains dried serial dilutions ofthree established antifungal agents and the colorimetric indicatorAlamar Blue in individual wells. Based on earlier multicentercomparisons of NCCLS and YeastOne colorimetric MIC data (6,12), the YeastOne plate was recently cleared by the Food andDrug Administration (FDA) for testing Candida spp. with fluconazole,itraconazole, and flucytosine in the clinical laboratory.

The purpose of this study was to extend the evaluation of theperformance of the YeastOne plate for testing four new antifungalagents: caspofungin acetate, posaconazole, ravuconazole, andvoriconazole. The evaluation was performed in three independentlaboratories on a set of 100 isolates of Candida spp. Two establishedantifungal agents (amphotericin B and fluconazole) were includedas controls. Colorimetric MICs (at 24 and 48 h) of the six agentswere compared to reference NCCLS M27-A2 MICs for each isolate-drugcombination in each of the three centers. The interlaboratoryreproducibility of colorimetric and reference MICs was alsoexamined.

MATERIALS AND METHODS

Top

Materials and Methods

Study design. The study was designed to compare MICs obtained by the YeastOneplate with those obtained by the NCCLS reference M27-A2 brothmicrodilution method (9) in the three independent laboratories(each laboratory tested 100 isolates by each method with sixagents). Two MIC readings were performed by each method, e.g.,at 24 and 48 h. Each first and second day, colorimetric MICsfor each isolate-drug combination were compared to both correspondingfirst- and second-day M27-A2 MICs. In addition, the interlaboratoryreproducibility of the colorimetric and reference MIC resultswas determined by performing repetitive testing (three timeson three different days in each center) with the two qualitycontrol [QC] strains.

Clinical isolates. A total of 100 isolates from the culture collection of the Universityof Iowa College of Medicine included 38 Candida albicans, 24C. glabrata, 7 C. krusei, 5 C. lusitaniae, 10 C. parapsilosis,and 16 C. tropicalis isolates. The NCCLS QC isolates, C. kruseiATCC 6258 and C. parapsilosis ATCC 22019, were also tested eachtime a set of clinical isolates was evaluated by the two procedures.The isolates were recovered from oral cavities, blood, or otherbody fluids and included strains with different susceptibilitypatterns to fluconazole (resistant, susceptible dose-dependent,and susceptible isolates). Each isolate represented a uniquestrain from a single patient. The isolates were maintained insterile water and subcultured onto antimicrobial-free mediumto ensure viability and purity prior to testing.

Antifungal agents. The YeastOne plates and reference microdilution plates, containingserial drug dilutions of caspofungin (Merck Research Laboratories,Rahway, N.J.), posaconazole (Schering-Plough Research Institute,Kenilworth, N.J.), ravuconazole (Bristol-Myers Squibb PharmaceuticalResearch Institute, Wallingford, Conn.), voriconazole (PfizerPharmaceuticals, New York, N.Y.), and the established agentsamphotericin B (Bristol-Myers Squibb Pharmaceutical ResearchInstitute) and fluconazole (Pfizer Pharmaceuticals), were providedby TREK. Drug dilutions of amphotericin B, caspofungin, andthe new triazoles ranged from 16 to 0.008 µg/ml, and fluconazoleconcentrations ranged from 128 to 0.12 µg/ml in both thereference and YeastOne plates. YeastOne plates, which are individuallypacked in foil and silica gel desiccant, were stored at ambienttemperature until testing was performed. The reference microdilutionplates were prepared following the NCCLS M27-A2 procedure (9),shipped frozen to each laboratory, and stored at -70°C untilthe day of the test.

Inoculum preparation. Stock inoculum suspensions of the yeasts were obtained from24-h cultures on Sabouraud dextrose agar at 35°C. The turbidityof each yeast suspension was adjusted by the spectrophotometricmethod (9).

Sensititre YeastOne Colorimetric Antifungal plate procedure. On the day of the test, the dried YeastOne plates were rehydratedwith the working yeast suspension (approximately 1.5 x 10³ CFU/ml)by dispensing 100 µl into each well. The YeastOne plateswere sealed and incubated at 35°C in a non-CO₂ incubatorand were read after 24 and 48 h of incubation by means of aview box. Colorimetric MICs were interpreted as the lowest concentrationof antifungal solutions that remained blue or unchanged (indicatingno growth) (amphotericin B and caspofungin) or the first dilutionthat either changed from blue to slightly purple (which is equivalentto the prominent growth inhibition [ $~$ 50%] seen by the NCCLS method)or remained unchanged (triazoles). QC isolates were tested inthe same manner by each participating laboratory.

NCCLS broth microdilution method (M27-A2). Following inoculation of the reference microdilution plates,plates were incubated at 35°C in a non-CO₂ incubator, andMICs were determined after 24 and 48 h. Reference MICs correspondedto the lowest drug dilution that showed complete growth inhibition(caspofungin and amphotericin B) and the lowest drug dilutionthat showed prominent (50% or more) growth inhibition (the triazoles).QC isolates were tested in the same manner by each participatinglaboratory.

Data analysis. Both on-scale (e.g., 0.12 and 128 µg/ml) and off-scale(e.g., <0.12 and >128 µg/ml) MICs were includedin the analysis. For the comparison between the two methods,MICs for each drug-organism combination by each method obtainedin the three laboratories were compared as follows: (i) 24-hMIC pairs by both methods and (ii) 24-h colorimetric MICs versusM27 48-h MICs. For the reproducibility evaluation (agreementamong the centers of colorimetric and reference MICs), combinedsite modal MICs were determined for each drug-organism combinationat both incubation times. Agreement was calculated by comparingeach result to the modal value, and the number of MICs within±1 log₂ dilution was then used to obtain the percentagesof agreement among the three sites.

RESULTS AND DISCUSSION

Top

Results and Discussion

Table 1 lists the established MIC ranges for the two QC isolates(1, 9) and the percentages of agreement of colorimetric MICswith reference MIC ranges at both incubation times. ColorimetricMICs for the C. krusei ATCC 6258 isolate were within the establishedrange at 24 h, but two caspofungin and posaconazole MICs wereone dilution higher at 48 h than the upper limit establishedby the NCCLS. Although similar results were observed for theC. parapsilosis ATCC 22019 strain and MICs were clustered withina three- to two-dilution range with four of the six agents,one fluconazole and four ravuconazole colorimetric MICs werehigher than the reference range at 24 h. Therefore, it appearsthat different fluconazole and ravuconazole MIC ranges shouldbe established for the QC isolate C. parapsilosis ATCC 22019for testing by the YeastOne method. In the evaluation of YeastOneperformance with the two QC isolates by Pfaller et al. (12),fluconazole MICs for C. parapsilosis ATCC 22019 were withinthe reference range at both incubation times, but optimal agreementwith the NCCLS range was achieved only at 48 h for amphotericinB. In the present study, the amphotericin B colorimetric resultswere not dependent on the time of incubation (Table 1). Thereis no reasonable explanation for these contrasting results;however, they underscore the influence of the incubation timeon MIC data. Reproducibility results for the two QC isolatesconfirm the superior reliability of 24-h YeastOne MIC data forthe established agents and extend previous reports for the fournew antifungal agents tested, caspofungin, posaconazole, ravuconazole,and voriconazole.

View this table:
[in this window]
[in a new window]
TABLE 1. Percent agreement of YeastOne colorimetric MICs of six antifungal agents with reference MIC ranges^a

The most important role of antifungal susceptibility testingis to identify resistant isolates among susceptible strains,which requires wide MIC ranges. Table 2 summarizes NCCLS MICsfor the actual number of isolates per species tested by bothmethods, as well as the percent agreement between YeastOne MICsand those reference values at the listed incubation times. YeastOneMICs had the same range at 24 h as the reference MICs, whichindicated that the colorimetric method could identify fluconazoleresistance ( $>=$ 64 µg/ml) and other isolatesfor which high reference MICs were obtained with the other agents.Although two major errors were observed with 24-h fluconazolecolorimetric results (0.5 and 4 µg/ml versus 64 µg/mlby the reference method), they represented the low reproducibilityamong the laboratories for two isolates (only one of the threecenters reported the low values). However, minor errors wereobserved among C. albicans (7 errors, 91% categorical agreement)and C. glabrata (29 errors, 54% categorical agreement). Forthe other species, the categorical agreement was 100%. Interpretivebreakpoints have not been established for amphotericin B, caspofungin,or the new triazoles against any fungal species. However, substantialdisagreement between the methods was low for those agents (0.3to 1%) when colorimetric MICs were read at 24 h (e.g., referenceMICs of $<=$ 1 µg/ml versus colorimetric MICs of $>=$ 8µg/ml). In general, results for established and new agentswere similar to those previously reported (2, 3, 5, 8, 10, 11),but the higher MICs ( $>=$ 2 µg/ml) of amphotericinB recently reported for C. krusei and C. glabrata (11) wereobtained mostly at 48 h.

View this table:
[in this window]
[in a new window]
TABLE 2. NCCLS broth microdilution MICs as reported by three laboratories for 100 clinical isolates of Candida spp. and percent agreement between YeastOne colorimetric and NCCLS MICs

The percent agreement between the M27-A2 and YeastOne panelresults was dependent on species, antifungal agent, and, especially,length of incubation for five of six agents (Table 2). For amphotericinB, the overall agreement was similar at both 24 and 48 h (97to 98%). In the earlier evaluation with >1,000 yeast isolates,the 24-h amphotericin B MICs by the YeastOne plate providedthe lowest percentages of agreement between the methods (40to 76% versus 90 to 99% at 48 h) (6). However, that evaluationincluded uncommon Candida spp., Cryptococcus spp., and otheryeasts and yeast-like organisms, and these were the specieswhere the major discrepancies were noted when this agent wastested. In the present study, we tested only the common andless fastidious Candida species for which the YeastOne platehas been cleared for clinical testing by the FDA. For the otheragents, the overall agreement between the two methods for colorimetricMICs (all the species tested) was better at 24 (95 to 99%) thanat 48 h (80 to 97%). The main reason for this discrepancy wasthat colorimetric 48-h MICs were consistently more than twodilutions higher than reference MICs, especially for isolatesof C. albicans, C. glabrata, and C. tropicalis. The agreementbetween the methods for C. krusei and C. parapsilosis was lessdependent on the incubation time, with 93 to 100% agreementfor all comparisons.

The interlaboratory agreement among the three centers for thecolorimetric YeastOne MICs was excellent (96 to 99%) at 24 h;lower percentages of interlaboratory agreement were observedfor ravuconazole (83%) and voriconazole (88%) at 48 h (Table3). The reproducibility of YeastOne MICs at 24 h was comparableto that obtained with reference MIC endpoints (95 to 99%). Reproducibilitywas lower at 48 h for both methods, but the effect of the longerincubation was more evident for colorimetric MICs (83 to 99%)than for the reference MICs (90 to 99%). It is noteworthy thatthe shorter incubation time had yielded superior interlaboratoryagreement for fluconazole NCCLS MICs in an earlier study (4).Data from the present study corroborated our earlier resultsregarding the YeastOne performance for clinical and QC isolates(6, 12).

View this table:
[in this window]
[in a new window]
TABLE 3. Percent interlaboratory agreement of colorimetric and NCCLS broth microdilution MICs for 100 Candida isolates^a

There are two commercially available antifungal susceptibilitytesting systems for fungal pathogens in the United States, theEtest (7) and the YeastOne plate (6, 12), but only the YeastOneplate has been cleared by the FDA for testing three establishedantifungal agents. Our evaluation of the performance of theSensititre YeastOne Colorimetric Antifungal plate suggests itspotential value also for use in the clinical laboratory forthe antifungal susceptibility testing of most Candida spp. after24 h of incubation with the new antifungal agents caspofungin,posaconazole, ravuconazole, and voriconazole. The clinical valueof these in vitro results needs to be established.

ACKNOWLEDGMENTS

This study was partially supported by a grant from TREK DiagnosticSystems.

FOOTNOTES

* Corresponding author. Mailing Address: Medical College of Virginia of Virginia Commonwealth University, P. O. Box 980049, Richmond, VA 23298-0049. Phone: (804) 828-9711. Fax: (804) 828-3097. E-mail: avingroff{at}hsc.vcu.edu.

REFERENCES

Top