In Vitro Susceptibility Testing of Geotrichum capitatum: Comparison of the E-Test, Disk Diffusion, and Sensititre Colorimetric Methods with the NCCLS M27-A2 Broth Microdilution Reference Method

Abstract

The in vitro activities of amphotericin B, flucytosine, fluconazole, itraconazole, and voriconazole against 23 isolates of Geotrichum capitatum were determined by the National Committee for Clinical Laboratory Standards (NCCLS) M27-A2 microdilution method and the Sensititre and agar diffusion methods. Amphotericin B and voriconazole appeared to be the more active drugs. Sensititre showed the highest rates of agreement with the NCCLS M27-A2 method.

Geotrichum capitatum, formerly known as Trichosporon capitatum or Blastoschizomyces capitatus, is an uncommon but frequently fatal cause of invasive infections in immunocompromised patients, particularly those with hematological malignancies (1, 3, 5, 6). Data on its antifungal agent susceptibilities are limited (3, 4, 5, 14); however, fluconazole-resistant strains have been reported (3, 4). The National Committee for Clinical Laboratory Standards (NCCLS) provides no specific guidelines for in vitro antifungal susceptibility testing for this fungus. Other susceptibility methods also have not been validated for use with G. capitatum. In this study, we evaluated the in vitro activities of five antifungal agents, including voriconazole, against clinical isolates of G. capitatum, using the standard broth microdilution NCCLS M27-A2 method (9) and alternative methods.

The study was conducted on 23 clinical isolates of G. capitatum collected from blood (n = 17), lower respiratory tract (n = 2), upper respiratory tract (n = 2), and urinary tract (n = 2) specimens. Quality control strains of Candida albicans (ATCC 90028) and Candida parapsilosis (ATCC 22019) were included in every test run. In vitro susceptibilities to amphotericin B, flucytosine, itraconazole, fluconazole, and voriconazole were determined by the NCCLS M27-A2 broth microdilution method (9), Sensititre YeastOne colorimetric method (AccuMedInternational) (8), and E-test method (12). Susceptibilities to fluconazole and voriconazole were also evaluated by the NCCLS M44-P disk diffusion method (7, 10, 12). Both agar diffusion methods were used with three different agars: RPMI 1640 with 2% glucose and 1.5% Bacto Agar (Difco Laboratories; RPMI-2%), Mueller-Hinton agar (Difco Laboratories) supplemented with 2% glucose and methylene blue (0.5 μg/ml; MHB), and Antibiotic Medium 3 with 2% glucose and 1.5% Bacto Agar (Difco Laboratories; AM3-2%). MIC endpoints were determined after 24 (Sensititre and disk diffusion) or 48 (NCCLS and E-test methods) h of incubation at 35°C. For disk diffusion results, the diameter of each zone of inhibition was taken as the area that showed a sharp decline (≥80%) in the density of growth (2). To facilitate reading more consistent and objective endpoints, the BIOMIC image analysis plate reader system (Giles Scientific, Santa Barbara, Calif.) was used (12), and these values were automatically converted into MICs by the BIOMIC System regression analysis software, which is calibrated against the NCCLS reference method MICs. Ranges for MICs and for MICs at which 50 and 90% of the isolates tested were inhibited were calculated for each of the methods used. For each testing method, the rate of agreement (within ±2 log₂ dilutions) with the NCCLS reference method was calculated. MICs determined by the E-test and disk diffusion methods that fell between the twofold dilutions of the MICs of the NCCLS method were elevated to the next drug concentration so that they matched the twofold dilution scheme. Isolates were classified as susceptible (S), susceptible-dose-dependent (S-DD) orintermediate (I), or resistant (R) to fluconazole, itraconazole, and flucytosine based on the criteria of Rex et al. (13) and those of the NCCLS (9). NCCLS-validated breakpoints have not yet been established for amphotericin B or voriconazole, but for the purposes of this study, we adapted the arbitrarily chosen criteria for these drugs proposed by Pfaller et al. (11, 12), who considered MICs of ≤1 μg/ml (for both drugs) indicative of susceptibility.

The interpretive classifications of the isolates yielded by each test method were compared to those obtained with the NCCLS reference method, and discrepancies were defined as follows: (i) minor, isolate classified as S or R by one test (reference method or alternative method) and as S-DD or I by the other method; (ii) major, isolate classified as S by the reference method and R by the other method; and (iii) very major, isolate classified as R by the reference method and S by the other method.

The MICs for the two quality control Candida strains yielded by the various methods were all within the reference ranges, with the exception of those that emerged for flucytosine in E-tests with MHB agar, both of which were high (>32 μg/ml).

In G. capitatum susceptibility testing with the NCCLS M27-A2 method, fungal growth was clearly visible after 24 h of incubation and somewhat more evident by 48 h, but the 24- and 48-h MICs never differed by more than one dilution. Trailing growth, which did not occur with amphotericin B, was observed with most strains in the presence of fluconazole and itraconazole and occasionally with flucytosine and voriconazole. Nonetheless, interpretation of MICs was not difficult.

Tables 1 and 2 summarize the antifungal susceptibility profiles of the 23 G. capitatum isolates. A review of reference method data in these two tables shows that all G. capitatum isolates were susceptible to both amphotericin B and voriconazole, and 90% were inhibited by very low concentrations of the drugs (0.125 and 0.25 μg/ml, respectively). For these two drugs, the other methods showed good concordance with the reference method. The only exception was the disk diffusion method on MHB agar for voriconazole.

TABLE 1.

In vitro susceptibilities of 23 isolates of G. capitatum to five antifungal agents as determined by different methods

Antifungal agent	Test method and medium	MIC (μg/ml)a			% MIC agreementb
		Range	50%	90%
Amphotericin B	NCCLS M27-A	0.06-0.25	0.125	0.125
	Sensititre	0.016-0.5	0.25	0.5	100
	E-test RPMI-2%	0.016-0.25	0.125	0.25	100
	E-test MHB	0.016-1	0.5	1	78.3
	E-test AM3-2%	0.016-0.25	0.25	0.25	95.7
Flucytosine	NCCLS M27-A	0.125-16	0.125	4
	Sensititre	0.03-16	0.03	8	95.7
	E-test RPMI-2%	0.016->32	0.03	>32	69.6
	E-test MHB	>32	>32	>32	8.7
	E-test AM3-2%	0.125->32	0.5	>32	78.3
Itraconazole	NCCLS M27-A	0.03-0.5	0.125	0.25
	Sensititre	2-6	0.25	0.5	100
	E-test RPMI-2%	0.125-1	0.5	1	73.9
	E-test MHB	0.125-2	0.5	1	43.5
	E-test AM3-2%	0.016-2	1	1	56.5
Fluconazole	NCCLS M27-A	1-32	8	8
	Sensititre	1-64	4	32	91.3
	E-test RPMI-2%	1-64	4	32	87
	E-test MHB	2-64	8	64	78.3
	E-test AM3-2%	0.25-32	4	8	87
	Disk RPMI-2%	0.06-64	4	16	87
	Disk MHB	2-64	16	64	82.6
	Disk AM3-2%	1-64	4	16	87
Voriconazole	NCCLS M27-A	0.03-0.5	0.25	0.25
	Sensititre	0.016-0.25	0.125	0.25	91.3
	E-test RPMI-2%	0.06-0.5	0.125	0.25	100
	E-test MHB	0.06-0.5	0.25	0.25	100
	E-test AM3-2%	0.016-0.125	0.125	0.125	95.7
	Disk RPMI-2%	0.06-1	0.25	1	100
	Disk MHB	0.06-4	1	4	47.8
	Disk AM3-2%	0.016-1	0.25	1	91.3

^a50% and 90%, MICs that inhibited 50% and 90% of the isolates tested, respectively.

^bPercentage of Sensititre, E-test, and disk diffusion MICs that were within 2 log₂ dilutions of the reference method (NCCLS M27-A) MICs.

TABLE 2.

Interpretive antifungal susceptibility classification of 23 isolates of G. capitatum determined by different methods

Antifungal agent	Method and medium	% of isolates by categorya			% of discrepancy resultsb			% Categorical agreementc
		S	S-DD and/or I	R	Minor	Major	Very major
Amphotericin B	NCCLS M27-A	100	0	0
	Sensititre	100	0	0	0	0	0	100
	E-test RPMI-2%	100	0	0	0	0	0	100
	E-test MHB	100	0	0	0	0	0	100
	E-test AM3-2%	100	0	0	0	0	0	100
Flucytosine	NCCLS M27-A	91.3	8.7	0
	Sensititre	82.6	17.4	0	8.7	0	0	91.3
	E-test RPMI-2%	82.6	0	17.4	8.7	8.7	0	82.6
	E-test MHB	0	0	100	8.7	91.3	0	0
	E-test AM3-2%	78.3	0	21.7	8.7	13	0	78.3
Itraconazole	NCCLS M27-A	65.2	34.8	0
	Sensititre	43.5	56.5	0	21.6	0	0	78.3
	E-test RPMI-2%	17.4	47.8	34.8	52.2	17.4	0	30.4
	E-test MHB	4.3	47.8	47.8	47.8	30.4	0	21.7
	E-test AM3-2%	17.4	26.1	56.5	26	39.1	0	34.8
Fluconazole	NCCLS M27-A	91.3	8.7	0			0
	Sensititre	82.6	13	4.3	13	0	0	87
	E-test RPMI-2%	78.3	8.7	13	8.7	4.3	0	87
	E-test MHB	65.2	13	21.7	21.7	13	0	65.3
	E-test AM3-2%	91.3	8.7	0	8.7	0	0	91.3
	Disk RPMI-2%	82.6	13	4.3	13	0	0	87
	Disk MHB	47.8	30.4	21.7	43.5	13	0	43.5
	Disk AM3-2%	87	8.7	4.3	17.4	0	0	82.6
Voriconazole	NCCLS M27-A	100	0	0
	Sensititre	100	0	0	0	0	0	100
	E-test RPMI-2%	100	0	0	0	0	0	100
	E-test MHB	100	0	0	0	0	0	100
	E-test AM3-2%	100	0	0	0	0	0	100
	Disk RPMI-2%	100	0		0	0	0	100
	Disk MHB	56.5	0	43.5	0	43.5	0	56.5
	Disk AM3-2%	100	0	0	0	0	0	100

^aPercentage of isolates classified in the given category. See Materials and Methods for definitions.

^bPercentage of results representing minor, major, or very major discrepancies with respect to those of the reference (NCCLS M27-A) method. See Materials and Methods for definitions.

^cAgreement rates reflect the percentage of isolates classified in the same category by the reference (NCCLS M27-A) method and alternative method.

The reference method MICs of flucytosine, itraconazole, and fluconazole tended to be clustered within the upper limits of the S category and in the S-DD and/or I category. Sensititre results showed the highest rates of agreement with those of the NCCLS M27-A2 method. Regardless of the agar used, E-test results for these three drugs showed lower concordance with the reference method data. The poorest results were seen when flucytosine was subjected to the E-test on MHB agar. When discrepancies were observed between E-test and reference method results for these three drugs, the MICs indicated by the E-test were generally higher. Disk diffusion test results for fluconazole showed good concordance with the reference method data, with the exception of the test on MHB agar, which showed very low rates of isolate categorical agreement (43.5%).

As shown in Table 2, none of the interpretive classifications yielded by the alternative methods represented “very major” discrepancies with respect to the NCCLS reference method. Concordance was particularly high for data for amphotericin B (100% agreement) and voriconazole. For the latter drug, disk diffusion testing on MHB agar was the only method that yielded discordant data (major discrepancies for 43.5% of all strains tested).

In summary, this extensive analysis of methods currently available for determination of the in vitro activities of amphotericin B, flucytosine, itraconazole, fluconazole, and voriconazole provides antifungal susceptibility data for the largest series of G. capitatum isolates tested thus far in a single setting. We did not encounter difficulties in the interpretation of the NCCLS M27-A2 method results. As alternatives to this method, the Sensititre Yeast One method proved to be the most “reliable” for evaluation of all five antifungal agents, although acceptable results were also obtained with E-tests on RPMI-2% or AM3-2% agar. For fluconazole and voriconazole susceptibility testing, reliable results were obtained with all three of the E-test methods and with disk diffusion testing (except that performed on MHB agar). Our findings tend to confirm previous observations on the high activity of amphotericin B against G. capitatum (6, 14) and the reduced susceptibility of some strains to flucytosine, fluconazole, and itraconazole (3). They also provide encouraging evidence of the potential role of voriconazole in our therapeutic armamentarium against invasive G. capitatum infections.