Evaluation of Etest Method for Determining Fluconazole and Voriconazole MICs for 279 Clinical Isolates of Candida Species Infrequently Isolated from Blood

The performance of Etest in fluconazole and voriconazole testingof 279 isolates of uncommon Candida spp. was assessed in comparisonwith the National Committee for Clinical Laboratory Standards(NCCLS)-approved standard broth microdilution (BMD) method.The NCCLS method employed RPMI 1640 broth medium, and MICs wereread after incubation for 48 h at 35°C. Etest MICs weredetermined with RPMI agar containing 2% glucose and were readafter incubation for 48 h at 35°C. The isolates includeCandida krusei, C. lusitaniae, C. guilliermondii, C. kefyr,C. rugosa, C. lipolytica, C. pelliculosa, C. dubliniensis, C.famata, C. zeylanoides, C. inconspicua, and C. norvegensis.Overall agreement between Etest and BMD MICs was 96% for fluconazoleand 95% for voriconazole. Where a discrepancy was observed betweenEtest and the reference method, the Etest tended to give lowervalues with both fluconazole and voriconazole. The Etest methodusing RPMI agar appears to be a useful method for determiningfluconazole and voriconazole susceptibilities of uncommon speciesof Candida.

INTRODUCTION

Top

Introduction

The Etest stable agar gradient MIC method (AB BIODISK, Solna,Sweden) has been shown to be useful in testing Candida spp.against a variety of antifungal agents, including fluconazoleand voriconazole (1, 3, 9, 11, 12, 14, 16, 17, 22, 24). Thespecies of Candida tested in these studies generally representthose most commonly isolated from clinical sources and are dominatedby Candida albicans, C. glabrata, C. parapsilosis, and C. tropicalis,which account for 95 to 97% of all clinical isolates of Candidaspp. (8, 13, 15, 18). Thus, although Etest has been validatedfor the four most common species of Candida, the evidence supportingits use in testing the less common species is lacking.

Among the approximately 17 species of Candida reported to causebloodstream infections (BSI) (8), 12 or 13 of these speciesaccount for less than 5% of all Candida BSI (19). These rarespecies include among others, C. krusei, C. lusitaniae, C. guilliermondii,C. kefyr, C. rugosa, and C. dubliniensis, several of which maypose problems with antifungal resistance and nosocomial spread(4, 5, 7, 23, 26). Although less common than C. albicans, C.glabrata, C. parapsilosis, and C. tropicalis, these speciesmay pose difficult management problems for individual patients,which may benefit from the application of antifungal susceptibilitytesting (21). Given the use of Etest for antifungal susceptibilitytesting of the common Candida spp. causing BSI, it is reasonableto validate its use for testing systemically active agents,such as fluconazole and voriconazole, against these less commonspecies as well.

The purpose of the present study is to expand the Etest databasefor fluconazole and voriconazole by testing an internationalcollection of 279 clinical BSI isolates of 12 uncommon speciesof Candida obtained from 68 different locations in 26 nations.The fluconazole and voriconazole MICs determined by Etest arecompared to MICs determined by the National Committee for ClinicalLaboratory Standards (NCCLS) reference broth microdilution (BMD)method, NCCLS M27-A (10).

MATERIALS AND METHODS

Top

Materials and Methods

Organisms. A total of 279 clinical isolates of Candida spp. obtained from68 medical centers in North America (39 centers), Latin America(6 centers), Europe (15 centers), and the Asia-Pacific region(8 centers) were tested. The collection included the followingnumbers of isolates: 118, C. krusei; 56, C. lusitaniae; 53,C. guilliermondii; 11, C. kefyr; 10, C. rugosa; 8, C. lipolytica;8, C. pelliculosa; 7, C. dubliniensis; 3, C. famata; 3, C. zeylanoides;1, C. inconspicua; and 1, C. norvegensis. All were incidentisolates obtained from blood cultures of 279 different patientswith candidemia. Isolates were identified by using Vitek andAPI yeast identification systems (bioMerieux, Inc., Hazelwood,Mo.) and were supplemented by conventional methods as needed(25). Isolates were stored as water suspensions until used.Prior to testing, each isolate was passaged at least twice onpotato dextrose agar (Remel, Lenexa, Kans.) to ensure purityand viability.

Susceptibility testing. Reference antifungal susceptibility testing of Candida spp.was performed by BMD, as described by the NCCLS (10). Referencepowders of fluconazole and voriconazole were obtained from PfizerPharmaceuticals (Groton, Conn.).

Etest strips for fluconazole and voriconazole were providedby AB BIODISK. MICs using Etest were determined as describedpreviously (11, 14) by using 150-mm-diameter plates containingRPMI agar with 2% glucose (RPG; Remel), an inoculum suspensionadjusted to the turbidity of a 0.5 McFarland standard (10⁶ cells/ml),and incubation at 35°C for 48 h. Both fluconazole and voriconazolestrips were placed on the same plate. The MICs of both fluconazoleand voriconazole were read at the lowest concentration at whichthe border of the elliptical inhibition zone intercepted thescale on the strip. Any growth, such as microcolonies, throughouta discernible inhibition ellipse was ignored.

MIC interpretive criteria for fluconazole were those publishedby Rex et al. (20) and the NCCLS (10). Breakpoints were as follows:susceptible (S), $<=$ 8 µg/ml; susceptible-dose dependent,16 to 32 µg/ml; and resistant (R), $>=$ 64 µg/ml. Interpretivebreakpoints have not yet been established for voriconazole.

QC. Quality control (QC) was performed for BMD and Etest in accordancewith NCCLS document M27-A by using C. krusei ATCC 6258 and C.parapsilosis ATCC 22019 (2, 10). QC determinations made on eachday of testing were within the control limits for fluconazoleand voriconazole described by Barry et al. (2, 3).

Analysis of results. Etest MICs for fluconazole and voriconazole read at 48 h werecompared to reference BMD MICs read at 48 h. The reference microdilutionMICs and Etest MICs were determined in two different time periodsand were read independently by two different individuals; i.e.,the testing was blinded. Etest MICs were rounded to the nexteven log₂ concentration in order to simplify analysis (3, 11,14). Discrepancies of no more than 2 dilutions were used tocalculate the percent agreement.

The interpretive breakpoints described by NCCLS (10) were utilizedto determine the categorical agreement between the Etest andBMD results for fluconazole. Major errors were identified asa classification of R by Etest and S by BMD, very major errorswere when the Etest results was S and BMD was R, and minor errorswere when one of the test results was S or R and the other wassusceptible-dose dependent.

RESULTS AND DISCUSSION

Top

Results and Discussion

Tables 1 and 2 summarize the in vitro susceptibilities of 279BSI isolates of uncommon Candida spp. to fluconazole and voriconazole,respectively, as determined by the reference BMD and Etest methods.The data are presented in a continuous fashion as cumulativepercentages of organisms susceptible at each dilution throughoutthe dilution series. The fluconazole MICs obtained by both BMDand Etest demonstrate excellent activity ( $>=$ 95% S) against C.lusitaniae, C. kefyr, C. pelliculosa, C. dubliniensis, C. zeylanoides,and C. inconspicua and relatively poor activity against C. krusei,C. guilliermondii, C. rugosa, C. lipolytica, and C. norvegensis(Table 1).

View this table:
[in this window]
[in a new window]
TABLE 1. In vitro susceptibility of uncommon Candida spp. to fluconazole determined by BMD and Etest

View this table:
[in this window]
[in a new window]
TABLE 2. In vitro susceptibility of uncommon Candida spp. to voriconazole determined by BMD and Etest

The MICs obtained by both BMD and Etest methods demonstratedthat voriconazole was very active against all of the uncommonspecies of Candida (92 to 100% S if MIC $<=$ 1 µg/ml) withthe exception of C. lipolytica (88% S at MIC if 1 µg/ml)(Table 2). Overall, 97 and 99% of isolates were inhibited at $<=$ 1µg/ml by BMD and Etest, respectively (Table 2).

As reported previously for the more common Candida spp. (3,11, 14), the agreement between BMD and Etest for both fluconazoleand voriconazole with the uncommon Candida spp. was excellent(Tables 1 and 2). The agreement was 96% for fluconazole and95% for voriconazole.

The agreement between fluconazole Etest and BMD MICs was >97% for all species, with the exception of C. guilliermondii(91%), C. famata (two of three, 67%), and C. norvegensis (zeroof one, 0%) (Table 1). When discrepancies were observed betweenthe results obtained by Etest and BMD for fluconazole, the Etestgenerally provided lower MICs, although Etest MICs tended tobe higher than BMD MICs if C. krusei was being tested. The overallcategorical agreement between Etest and BMD results for fluconazolewas 81%, with 1.1% very major errors, 0% major errors, and 17.9%minor errors (data not shown). This agreement is lower thanthat reported by Barry et al. (3) when the more common speciesof Candida are tested, largely due to a greater number of minorerrors attributed to C. krusei.

The agreement between voriconazole Etest and BMD MICs was >99% for all species, with the exception of C. guilliermondii(79%), C. pelliculosa (seven of eight, 88%), and C. famata (twoof three, 67%) (Table 2). As with fluconazole, when discrepanciesoccurred between the results obtained by Etest and BMD for voriconazole,the Etest provided lower MICs.

The results of this study confirm and extend those of previousreports regarding the ability of Etest to generate fluconazoleand voriconazole MIC data for the less common species of Candida(1, 6, 9, 11, 14). Previously, agreement was demonstrated betweenEtest (by using RPG) and the reference BMD method of 94% forfluconazole and 98% for voriconazole in studies where the vastmajority of isolates were the more common species of Candida(11, 14). Testing smaller numbers of C. krusei, C. lusitaniae,and C. guilliermondii isolates revealed agreements between Etestand BMD of 97 to 100% for fluconazole (11) and 100% for voriconazole(14). Favel et al. (6) reported an agreement between Etest andBMD of 92% when testing fluconazole against 35 isolates of C.lusitaniae.

In summary, we have provided documentation of the ability ofEtest to generate fluconazole and voriconazole MIC data foruncommon species of Candida that are comparable to those obtainedby the NCCLS BMD method. RPMI agar with 2% glucose may be usedto determine reference quality MICs of fluconazole and voriconazoleagainst these rare agents of candidemia as well as the morecommon species detected in the clinical laboratory. Althoughthe species tested in this study are uncommon causes of BSI,the fact that several may exhibit innate or acquired resistanceto both amphotericin B and fluconazole (4, 5, 7, 23, 26) emphasizesthe importance of reference quality antifungal testing capabilitiesin aiding management decisions (21). Given these concerns, thepotent in vitro activity of voriconazole against these speciesis notable (Table 2).

ACKNOWLEDGMENTS

Linda Elliott provided excellent support in the preparationof the manuscript.

The International Fungal Surveillance Program was supportedin part by research grants from Bristol-Myers Squibb (SENTRYProgram), Pfizer (ARTEMIS Program), and Schering Plough (GlobalSurveillance Program).

We express our appreciation to all International Fungal SurveillanceProgram participants: L. Steele-Moore, Christiana Care HealthServices, Wilmington, Del.; G. Denys, Clarion Health MethodistHospital, Indianapolis, Ind.; C. Staley, Henry Ford Hospital,Detroit, Mich.; J. R. Dipersio, Summa Health System, Akron,Ohio; M. Saubolle, Good Samaritan Regional Medical Center, Phoenix,Ariz.; M. L. Wilson, Denver General Hospital, Denver, Colo.;G. D. Overturf, University of New Mexico Hospital, Albuquerque;L. R. Peterson, Northwestern Memorial Hospital, Chicago, Ill.;P. C. Schreckenberger, University of Illinois at Chicago, Chicago;G. V. Doern, University of Iowa Hospitals and Clinics, IowaCity; et al.

FOOTNOTES

* Corresponding author. Mailing address: Medical Microbiology Division, C606 GH, Department of Pathology, University of Iowa College of Medicine Iowa City, IA 52242. Phone: (319) 384-9566. Fax: (319) 356-4916. E-mail: michael-pfaller{at}uiowa.edu.

Members of the International Fungal Surveillance ParticipantGroup are listed in Acknowledgments.

REFERENCES

Top