Correlation between E-Test, Disk Diffusion, and Microdilution Methods for Antifungal Susceptibility Testing of Fluconazole and Voriconazole

The activities of fluconazole and voriconazole against isolatesof Candida spp. (n = 400) were tested by the E-test, disk diffusion,and the National Committee for Clinical Laboratory Standards(NCCLS) M27-A2 broth microdilution-based reference methods.More than 96% of isolates found to be susceptible to fluconazoleby the reference method were identified as susceptible by theagar-based methods. Lesser degrees of correlation with the referencemethod were seen for isolates identified as resistant by theagar-based methods. Interpretive categories are not availablefor voriconazole, but results qualitatively similar to thosefor fluconazole were seen. The agar-based E-test and disk diffusionmethods are reliable alternatives to the NCCLS M27-A2 referencemicrodilution method for isolates that test susceptible to fluconazole.

INTRODUCTION

Top

Introduction

The development of standardized antifungal susceptibility testingmethods has been the subject of numerous studies during thelast decade. Reference methods for yeasts (the National Committeefor Clinical Laboratory Standards [NCCLS] M27-A2 method) andmolds (the NCCLS M38-A method) are now available (16). Agar-basedsusceptibility testing methods have been a focus of interestfor many researchers and include the classical disk diffusion(DD) methods and the E-test (ET) method (3, 6-10, 13, 14, 16-18).Those tests are very attractive due to their simplicity, reproducibility,and lack of requirements for specialized equipment (11, 16).Recent studies have documented comparable results between thosemethods and the results of standard reference broth microdilution(MD) susceptibility testing (7, 11, 13).

In this study, we compared the NCCLS M27-A2 MD method with theET and DD methods for determination of the susceptibilitiesof 400 Candida species isolates to fluconazole and voriconazole.The ET and DD methods are well studied for fluconazole (3, 6-11,14), and this work extends their usage to include voriconazole.

MATERIALS AND METHODS

Top

Materials and Methods

Isolates. Four hundred bloodstream isolates of Candida species were randomlyselected for testing. These included 205 isolates of Candidaalbicans, 56 isolates of C. tropicalis, 39 isolates of C. glabrata,66 isolates of C. parapsilosis, 24 isolates of C. krusei, and10 isolates of other species. The isolates were identified withthe API 20C AUX system (Biomerieux Vitek, Hazelwood, Mo.) andwere subsequently stored in sterile distilled water at roomtemperature until susceptibility tests were performed. Eachisolate was subcultured at least twice on Sabouraud dextroseagar and incubated at 35°C prior to testing to ensure purityand optimal growth.

Inoculum suspensions. Yeast inoculum suspensions were prepared as described for theNCCLS M27-A2 method (12). The turbidity was measured with aspectrophotometer at 530 nm and was adjusted to match a 0.5McFarland density standard, resulting in a concentration of1 x 10⁶ to 5 x 10⁶ yeast cells/ml. This inoculum was used directlyfor inoculation of agar plates (see below) or was diluted asneeded for the MD procedure.

Antifungal agents. Antifungal research powders were supplied by Pfizer Inc. (PfizerPharmaceuticals Group, New York, N.Y.) and stored at -20°Cuntil they were used. ET strips were obtained from AB Biodisk(Solna, Sweden), with the drug concentrations ranging from 0.016to 256 µg/ml for fluconazole and 0.002 to 32 µg/mlfor voriconazole. Paper disks containing 1 µg of voriconazolewere manufactured by Remel, Inc. (Lanexa, Kans.). Paper diskscontaining 25 µg of fluconazole were manufactured by BectonDickinson Microbiology Systems (Cockeysville, Md.).

Media and susceptibility testing methods. Broth MD testing was done by the NCCLS M27-A2 MD method andwas performed in RPMI 1640 buffered to pH 7.0 with 0.165 M morpholinepropanesulfonicacid obtained from Sigma Chemical Co. (St. Louis, Mo.). Theantifungal agents were tested over final concentration rangesof 0.125 to 64 µg/ml for fluconazole and 0.015 to 16 µg/mlfor voriconazole. The plates were incubated at 35°C andread with a spectrophotometer at 570 nm after 24 and 48 h. TheMIC was defined as the lowest drug concentration that reducedgrowth by 50% compared with the growth of the drug-free controls.

The ET and DD methods were performed on Mueller-Hinton agarsupplemented with 2% glucose and 0.5 µg of methylene blue(MB; Harleco, Gibbstown, N.J.) per ml due to the ability ofthat medium to produce enhanced definition of growth margins(5). To prepare the medium, stock solutions of MB (5 mg/ml)and glucose (0.4 g/ml) were made in distilled water. A totalof 100 µl of stock MB was added to 100 ml of stock glucosesolution to make a stock solution of 0.4 g of glucose per mlplus 5 µg of MB per ml (GMB). The GMB stock solution wasfilter sterilized and stored at 4°C. Mueller-Hinton agarplates (diameter, 15 cm, with 60 ml of agar; Becton DickinsonMicrobiology Systems) were prepared by pouring 2.9 ml of theGMB stock solution on the plate and allowing it to absorb for4 to 6 h before inoculation.

The agar plates were inoculated by dipping a sterile cottonswab into the inoculum and evenly streaking the swab in threedirections over the entire surface of the plate. The plateswere allowed to dry for at least 15 min before the ET stripsand the disks were applied to the surface. The ET strips anddisks with fluconazole and voriconazole were applied onto eachinoculated plate, and the plates were incubated at 35°C,with readings taken after 24 and 48 h. Inhibitory zone diametersfor the disks and the MICs for the ET strips were measured atthe transitional point where growth abruptly decreased, as determinedby a marked reduction in colony size, number, and density.

Interpretive breakpoints for fluconazole for the ET and theM27-A2 MD methods follow those published as part of the M27-A2method: susceptible, $<=$ 8 µg/ml; susceptible-dose dependent,16 to 32 µg/ml; and resistant, $>=$ 64 µg/ml (12). Forthe DD method, zone diameters were interpreted on the basisof the work of Barry et al. (5), with zone diameters of $>=$ 19 mmindicating susceptibility, zone diameters of 15 to 18 mm indicatingsusceptible-dose dependent, and zone diameters of $<=$ 14 mm indicatingresistance.

Quality control isolates C. parapsilosis ATCC 22019 and C. kruseiATCC 6258 were included in all runs, and all results were withinpublished limits (4, 12).

RESULTS AND DISCUSSION

Top

Results and Discussion

Table 1 shows the drug MICs and zone diameters obtained by allthree methods after both 24 and 48 h of incubation. Overall,the ET method tended to give slightly higher fluconazole MICsthan the MD method at both time points, whereas the voriconazoleMICs by the ET and MD methods were similar at both time points.

View this table:
[in this window]
[in a new window]
TABLE 1. Susceptibilities of 400 Candida spp. to fluconazole and voriconazole as determined by three methods at two incubation times

As shown in Table 2, the overall levels of agreement betweenthe MICs obtained by the MD and ET methods at 24 and 48 h weregood for both fluconazole and voriconazole. For both drugs,the agreement between the ET MIC at 48 h and the reference (MD)MIC at 48 h was >93%. However, better correlations were notedfor the readings obtained at 24 h, when the percent agreementwas >98% for the two drugs. Disparate readings were generallyattributable to trailing growth: for isolates for which therewas a difference, the ET MICs tended to be lower at both timepoints and to be lower by the MD method at 24 h, but the MDMICs tended to be elevated at 48 h.

View this table:
[in this window]
[in a new window]
TABLE 2. Percentage of paired MD and ET MICs within 2 doubling dilutions

Comparisons of the results obtained by all three methods areshown in Table 3 by interpretive category for fluconazole. Isolatesthat tested susceptible by the DD or ET method at either timepoint had a $>=$ 96% likelihood of testing susceptible or susceptible-dosedependent by the MD method. The MICs for isolates testing resistantby the DD and ET method were almost always in the susceptible-dosedependent or resistant category by the MD method, whereas isolatestesting resistant by the MD method produced results by the agar-basedmethod ranging from susceptible to resistant. Thus, the correlationbetween the resistant categories for the MD method and the agar-basedmethods was $<=$ 83%.

View this table:
[in this window]
[in a new window]
TABLE 3. Comparison of interpretive categories for fluconazole and rates of interpretive agreement^a

Table 4 correlates the voriconazole MICs obtained by the MDmethod and the inhibition zone diameters obtained by the DDmethod after 24 and 48 h of incubation. The best correlationwas obtained with readings obtained at 24 h for both methods.After 48 h, trailing growth similar to that noted above forfluconazole tended to generate higher MICs.

View this table:
[in this window]
[in a new window]
TABLE 4. Correlation between results of MD and DD methods for voriconazole after 24 and 48 h of incubation

As noted by others (5, 11), use of Mueller-Hinton agar floodedwith GMB enhanced growth and simplified reading relative tothe MD method. In addition, the trailing phenomenon was lesspronounced by both agar-based methods. As others have shownthat performing the ET method on RPMI 1640 or Casitone (themedium suggested by the manufacturer) produces results comparableto those obtained by the reference MD method (5), we did notrepeat that work. Rather, we tested the isolates by the ET methodon GMB-supplemented Mueller-Hinton agar. The results obtainedby both the DD and the ET methods were in acceptable concordancewith those obtained by the MD method, with the exception ofthe recurring problem of discrepancies due to isolates thatshowed trailing growth.

In summary, the agar-based ET and DD methods are reliable alternativesto the NCCLS M27-A2 reference MD method for isolates that testsusceptible to fluconazole. However, the detection of resistanceby agar-based methods correlates poorly with the detection ofresistance by the reference NCCLS M27-A2 method. Specifically, $>=$ 90% of the isolates that tested resistant to fluconazole byan agar-based method tested susceptible-dose dependent or resistantby the reference MD method. Conversely, 30 to 50% of the isolatesthat tested resistant by the MD method appeared to be susceptiblewhen they were tested by agar-based methods. This differencewas principally due to trailing growth associated with the MDmethod. Prior work suggests that the results for isolates withsignificant trailing should be interpreted on the basis of thelower MIC observed at the earlier time point (1, 2, 12, 15).Our data thus suggest that the results for isolates that appearto be resistant by any method should be carefully reviewed andthat such isolates may merit repeat testing and/or testing byan alternative method. Although more work needs to be done withless susceptible isolates, the aggregate data suggest that agar-basedmethods appear to produce a more consistent in vitro-in vivocorrelation than the reference MD method by eliminating trailinggrowth from the equation. The lack of interpretive breakpointsfor voriconazole makes such comparisons impossible for thisnewer triazole; however, analysis of numeric MICs and the correspondingzone diameters for this compound suggests conclusions similarto those for fluconazole.

FOOTNOTES

* Corresponding author. Present address: Faculty of Medical Sciences, Lebanese University, P.O. Box 55564 Sin-El-Fil, Beirut, Lebanon. E-mail: madonne{at}inco.com.lb.

REFERENCES

Top