Multicenter Study of Epidemiological Cutoff Values and Detection of Resistance in Candida spp. to Anidulafungin, Caspofungin, and Micafungin Using the Sensititre YeastOne Colorimetric Method

Abstract

Neither breakpoints (BPs) nor epidemiological cutoff values (ECVs) have been established for Candida spp. with anidulafungin, caspofungin, and micafungin when using the Sensititre YeastOne (SYO) broth dilution colorimetric method. In addition, reference caspofungin MICs have so far proven to be unreliable. Candida species wild-type (WT) MIC distributions (for microorganisms in a species/drug combination with no detectable phenotypic resistance) were established for 6,007 Candida albicans, 186 C. dubliniensis, 3,188 C. glabrata complex, 119 C. guilliermondii, 493 C. krusei, 205 C. lusitaniae, 3,136 C. parapsilosis complex, and 1,016 C. tropicalis isolates. SYO MIC data gathered from 38 laboratories in Australia, Canada, Europe, Mexico, New Zealand, South Africa, and the United States were pooled to statistically define SYO ECVs. ECVs for anidulafungin, caspofungin, and micafungin encompassing ≥97.5% of the statistically modeled population were, respectively, 0.12, 0.25, and 0.06 μg/ml for C. albicans, 0.12, 0.25, and 0.03 μg/ml for C. glabrata complex, 4, 2, and 4 μg/ml for C. parapsilosis complex, 0.5, 0.25, and 0.06 μg/ml for C. tropicalis, 0.25, 1, and 0.25 μg/ml for C. krusei, 0.25, 1, and 0.12 μg/ml for C. lusitaniae, 4, 2, and 2 μg/ml for C. guilliermondii, and 0.25, 0.25, and 0.12 μg/ml for C. dubliniensis. Species-specific SYO ECVs for anidulafungin, caspofungin, and micafungin correctly classified 72 (88.9%), 74 (91.4%), 76 (93.8%), respectively, of 81 Candida isolates with identified fks mutations. SYO ECVs may aid in detecting non-WT isolates with reduced susceptibility to anidulafungin, micafungin, and especially caspofungin, since testing the susceptibilities of Candida spp. to caspofungin by reference methodologies is not recommended.

INTRODUCTION

Invasive infections caused by Candida, Aspergillus, and other fungi are increasing in incidence and prevalence, especially among immunocompromised patients and/or those with serious underlying diseases (1–3). The attributable mortality rate due to candidemia can be as high as 47%, depending on the patient population and age (1, 3). Three echinocandins (anidulafungin, caspofungin, and micafungin) have been licensed for intravenous treatment and prevention of invasive Candida infections (including candidemia) (4, 5). The Clinical and Laboratory Standards Institute (CLSI) has established standard conditions for testing the susceptibilities of Candida spp. to the three echinocandins (6) including species-specific breakpoints (BPs) for echinocandin MIC interpretation (7). The main role of the echinocandin species-specific BPs is to predict the clinical outcome of treatment with these agents. In contrast, the role of method-dependent species-specific epidemiological cutoff values (ECVs) is to aid in detecting potentially resistant or less susceptible isolates for those species for which BPs are not available. Candida isolates that are potentially echinocandin resistant or less susceptible often harbor amino acid substitutions in the Fks1p (and/or Fks2p in C. glabrata) gene. Such mutants have been associated with breakthrough candidiasis (8) or treatment failure (9). Significant interlaboratory variability in caspofungin modal MICs (wide ranges) precludes routine testing or reporting of caspofungin MICs for Candida spp. derived by CLSI methodology (10). The high degree of caspofungin MIC interlaboratory variation may potentially lead to incorrect categorization of susceptibility results. Examination of the European Committee on Antimicrobial Susceptibility Testing (EUCAST) MIC distributions for the four common Candida spp. from seven laboratories revealed the same problem (10). As a result, EUCAST has not established BPs for caspofungin (10–12). At this time, method-dependent multilaboratory CLSI-generated ECVs have been defined only for anidulafungin and micafungin and several species of Candida (13).

The Sensititre YeastOne (SYO) colorimetric yeast susceptibility test is a commercial broth microdilution method that produces MIC data for Candida spp.; this method is widely used in both clinical and research laboratories. A recent study demonstrated low caspofungin modal MIC variability among nine laboratories that used the SYO method for testing the susceptibility of Candida spp. to echinocandins (14). However, the SYO method recommends the use of CLSI interpretive criteria for interpretation of MIC results (15). As a result, species-specific ECVs for susceptibility testing by the SYO method need to be established, especially for caspofungin and Candida spp. Since the SYO panel also provides MICs for the other two echinocandins, SYO ECVs for anidulafungin and micafungin would be useful for laboratories using this method. The ECV is the highest endpoint of the MIC distribution of the wild-type (WT) population and is established using MIC distributions from multiple laboratories (at least 3 laboratories and 100 MICs/species/agent) (13, 16, 17). ECV surveillance may detect the emergence of in vitro resistance or distinguish between phenotypic WT isolates (isolates with no detectable phenotypic resistance) and non-WT isolates (isolates with mechanisms of resistance) (13, 16–18). The SYO MIC data from 38 laboratories used to define SYO ECVs in the present study are representative of the susceptibility of these species to caspofungin and to the other two echinocandins as determined by this method.

The purposes of the present study were (i) to pool echinocandin MIC data generated using the SYO broth microdilution method originating from ≥12 laboratories for eight Candida spp. (Candida albicans, C. dubliniensis, C. glabrata complex, C. guilliermondii, C. krusei, C. lusitaniae, C. parapsilosis complex, and C. tropicalis) in order to define the WT susceptibility MIC distributions of anidulafungin, caspofungin, and micafungin and (ii) to propose method-dependent SYO ECVs for each of these echinocandin/species combinations for which the number of SYO MICs was >100. MIC distributions comprising between 18 and 44 isolates of less prevalent species (C. famata, C. kefyr, C. lipolytica, C. metapsilosis, C. orthopsilosis, C. pelliculosa, and C. rugosa) were also documented. Echinocandin MICs for 18 to 6,007 isolates (species and agent dependent) were pooled from data generated in 38 independent laboratories (in Australia, Canada, Europe, Mexico, New Zealand, South Africa, and the United States). Since most of the isolates included in the study were not assessed for mechanisms of resistance, we evaluated our ECVs using SYO MIC data for individual well-characterized non-WT isolates (isolates harboring mechanisms of resistance or fks1 and fks2 gene mutations) (8, 19–24) and 14 WT isolates (isolates with no phenotypic resistance or fks gene mutations) in the same manner that triazole, micafungin, and anidulafungin CLSI ECVs were evaluated in previous CLSI ECV studies (13, 17).

MATERIALS AND METHODS

Isolates.

The isolates evaluated were recovered from blood cultures from patients with candidemia and other sterile sites (81.5%) as well as from other sites (18.5%). Antifungal susceptibility testing for each unique isolate (no serial isolates) was performed by the SYO broth microdilution method (dry-form colorimetric panel) following the manufacturer's instructions (15) at the following medical centers: VCU Medical Center, Richmond, VA; Hospital Universitario Central de Asturias, Oviedo, Spain; Unidad de Microbiología del Centro de Investigación, Valencia, Spain; University of Michigan Health System, Ann Arbor, MI; Clinical Mycology Reference Laboratory, Westmead Hospital, New South Wales, Australia; University of Pittsburgh Medical Center, Pittsburgh, PA; Stanford University Medical Center, Palo Alto, CA; Universidad Autonóma de Nuevo León, Monterrey, Nuevo León, Mexico; National Institute for Communicable Diseases, Johannesburg, South Africa; Fondazione IRCCS Ospedale Maggiore Policlinico, Milan, Italy; University of Utah Hospital, Salt Lake City, UT; National Mycology Reference Centre, Adelaide, Australia; Shands Hospital, Gainesville, FL; New York Presbyterian Hospital, Columbia University Medical Center, New York, NY; Public Health Ontario, Toronto, Canada; Centers for Disease Control and Prevention, Atlanta, GA; Clinical Microbiology Laboratory, Attikon Hospital, National and Kapodistrian University of Athens, Athens, Greece; Auckland City Hospital, Auckland, New Zealand; Hospital General Universitario Gregorio Maraňón, Universidad Complutense, Madrid, Spain; UFI 11/25, Universidad del País Vasco-Euskal Herriko Unibertsitatea, Bilbao, Spain; Hospital Universitario Puerta del Mar, Cádiz, Spain; Hospital de la Santa Creu i Sant Pau, Barcelona, Spain; The Johns Hopkins Hospital, Baltimore, MD; Mayo Clinic, Rochester, MN; Hospital Universitario Son Dureta, Palma de Mallorca, Spain; Hospital Donostia, San Sebastián, Spain; Hospital Universitario Puerta de Hierro Majadahonda, Spain; Hospital Universitario Ramón y Cajal, Madrid, Spain; Hospital Universitario de Salamanca, Salamanca, Spain; Universidad de Córdoba, Hospital General Universitario Reina Sofía, Córdoba, Spain; Hospital Clinic, Barcelona, Spain; Universitario de San Carlos, Madrid, Spain; Hospital Universitario La Fe, Valencia, Spain; Complejo Hospitalario Universitario de Santiago, Santiago de Compostela, Spain; Hospital Universitario Valle Hebrón, Barcelona, Spain; Universitario Lozano Blesa, Zaragoza, Spain; Hospital Universitario Virgen del Rocío, Seville, Spain; and Hospital Universitario Virgen De La Arrixaca, Murcia, Spain. These laboratories were coded 1 to 41, but three laboratories were excluded from the study for reasons discussed below; data used for the analyses were from the remaining 38 laboratories. Isolates were identified at each medical center using conventional methodologies (e.g., morphology on cornmeal-Tween 80 agar, growth at 45°C, API 20C Aux yeast identification system, Vitek yeast biochemical card, or matrix-assisted laser desorption ionization–time of flight mass spectrometry [MALDI-TOF MS] since 2010) or by phenotypic and/or molecular (1 to 100% depending on the laboratory; overall, ∼17.6%) identification as needed (e.g., when the phenotypic profiles were inconclusive and from the laboratories reporting data for the less prevalent species) (25).

Available SYO MIC data for each agent were pooled for 6,007 C. albicans, 186 C. dubliniensis, 3,188 C. glabrata complex (C. glabrata sensu stricto, C. nivariensis, and C. bracariensis), 119 C. guilliermondii, 493 C. krusei, 205 C. lusitaniae, 3,136 C. parapsilosis complex (C. parapsilosis sensu stricto, C. metapsilosis, and C. orthopsilosis), and 1,016 C. tropicalis isolates originating from between 12 and 32 independent laboratories, as well as for other less prevalent species from three or more laboratories (C. famata, C. kefyr, C. lipolytica, C. pelliculosa, and C. rugosa) (Tables 1 and 2). In addition, 44 isolates were identified as C. orthopsilosis, 22 isolates as C. metapsilosis (data shown in Table 2), seven isolates as C. nivariensis, and four isolates as C. bracariensis (data not shown in Table 2). One or both quality control (QC) isolates (C. parapsilosis ATCC 22019 and C. krusei ATCC 6258) were used by the participant laboratories (7, 15).

TABLE 1.

Pooled MIC distributions of anidulafungin, caspofungin, and micafungin for eight species of Candida from between 12 and 32 laboratories determined by the commercial Sensititre YeastOne broth microdilution method

Agent	Species	No. of labs	No. of isolates tested	No. of isolates with MIC (μg/ml) ofa:
				≤0.008	0.016	0.03	0.06	0.12	0.25	0.5	1	2	4	≥8
Anidulafungin	C. albicans	16	3,118		753	1,109	727	424	89	10	4	2
	C. glabrata complex	16	2,520		360	1,023	692	367	24	11	12	23	7	1
	C. krusei	19	408		40	136	115	103	9	2	2	1
	C. parapsilosis complex	28	2,468		7	10	10	59	89	301	1,153	781	45	13
	C. tropicalis	14	652		63	83	131	294	67	7	5			2
	C. dubliniensis	12	131		13	23	38	41	6	1				9
	C. guilliermondii	14	101		2			4	6	16	43	26	4
	C. lusitaniae	14	185		3	6	16	117	35	7	1
Caspofungin	C. albicans	32	6,007	21	230	1,772	2,322	1,227	352	59	13	3	1	7
	C. glabrata complex	23	3,188	6	11	217	1,096	1,190	497	104	24	14	1	28
	C. krusei	20	493			1	10	40	238	148	43	8	4	1
	C. parapsilosis complex	32	3,136	14	9	39	102	208	640	1,201	677	224	13	9
	C. tropicalis	23	1,016	8	52	313	344	199	70	19	3	1	1	6
	C. dubliniensis	15	186		3	28	76	43	20	5		2		9
	C. guilliermondii	15	118				2	17	33	36	16	9	3	2
	C. lusitaniae	18	205	2	3	13	26	52	55	36	13	4	1
Micafungin	C. albicans	26	4,862	2,632	1,724	389	61	21	15	8	2	7		3
	C. glabrata complex	20	2,763	676	1,573	399	37	18	12	8	7	8	10	15
	C. krusei	19	444	2	4	7	71	254	98	5		1	2
	C. parapsilosis complex	32	2,659	3	9	4	11	24	113	417	1,251	747	69	11
	C. tropicalis	22	983	40	252	503	155	19	4	5	1	1	1	2
	C. dubliniensis	12	123	10	38	40	21	5						9
	C. guilliermondii	15	119		1	2	2	4	21	45	27	14	3
	C. lusitaniae	15	191	4	9	24	107	36	6	5

^aThe highest number in each row (showing the most frequently obtained MIC or the mode) is in boldface.

TABLE 2.

Pooled MIC distributions of anidulafungin, caspofungin, and micafungin for seven less prevalent species of Candida from >3 laboratories determined by the commercial Sensititre YeastOne broth microdilution method

Agent	Species	No. of isolates tested	No. of isolates with MIC (μg/ml) ofa:
			≤0.008	0.016	0.03	0.06	0.12	0.25	0.5	1	2	>4
Anidulafungin	C. famata	44		3	7	8	13	4	2	1	5	1
	C. kefyr	27			4	8	12	3
	C. lipolytica	21		1	2	4	6	5	3
	C. metapsilosis	22				1	7	7	4	3
	C. orthopsilosis	24					1	5	8	10
	C. pelliculosa	18		6	8	2	1	1
	C. rugosa	30		3	7	6	11	1	2
Caspofungin	C. famata	35			2	4	4	14	5	4		2
	C. kefyr	28	1	2	9	10	3	2	1
	C. lipolytica	22	1	1	1	2	2	4	8	3
	C. metapsilosis	22				2	4	8	6	2
	C. orthopsilosis	44					6	13	14	7	4
	C. pelliculosa	21		1	2	9	8	1
	C. rugosa	30			1	4	4	7	10	3	1
Micafungin	C. famata	44	3	3	8	11	6	7	2	2	2
	C. kefyr	27			9	15	2	1
	C. lipolytica	25	1	4	1		3	7	6	3
	C. metapsilosis	21					4	8	7	2
	C. orthopsilosis	20				1	4	6	8	1
	C. pelliculosa	20		3	10	6	1
	C. rugosa	25	1	3	4	7	5	4	1

^aThe highest number in each row (showing the most frequently obtained MIC or the mode) is in boldface.

SYO MICs for 81 well-characterized non-WT strains (isolates tested for the presence of fks1 or fks2 gene mutations) (8, 19–24) and 14 WT strains (with no fks gene mutations) also were used to assess the ability of various SYO ECVs of anidulafungin, caspofungin, and micafungin to identify Candida non-WT isolates. SYO MICs for these isolates were obtained in six of the participant laboratories.

Antifungal susceptibility testing.

MICs were obtained at each center by following the SYO method (dry-form colorimetric panel; final inoculum concentrations that ranged from 1.5 × 10³ to 8 × 10³ CFU/ml and 24 h of incubation, unless there was insufficient color change in the drug-free control at 24 h); caspofungin and micafungin concentrations ranged from 0.008 to 8 μg/ml and those of anidulafungin from 0.015 to 8 μg/ml. MICs were the lowest echinocandin concentration that resulted in substantial inhibition of growth of the organism being tested as detected by visually observing the color change compared to the drug-free growth control well (15) or where “less intense color change” was evident (MCS Diagnostics). Each center included either one or both QC reference strains each time that a set of clinical isolates was evaluated by the SYO microdilution method (or once a week). SYO MIC limits for both QC isolates are the same as those listed in the CLSI M27-S4 document with one exception: the SYO micafungin MIC range for the QC strain C. krusei ATCC 6258 is one dilution lower than the CLSI range (0.06 to 0.25 versus 0.12 to 0.5 μg/ml, respectively) (15). Discrepant MICs (lower) for the QC strains were occasionally reported for the QC isolate C. parapsilosis ATCC 22019 with micafungin and for the QC strain C. krusei ATCC 6258 with anidulafungin, but MICs for the clinical isolates were not included in the analyses when such results were documented. Therefore, MICs were pooled or used in the analysis only when MICs for the QC isolates were within the established MIC limits as listed by the manufacturer (15).

Definitions.

The definition of the ECV as well as of the two populations (WT and non-WT MIC populations/isolates) have been provided in detail elsewhere (13, 16–18). Briefly, a non-WT organism shows reduced susceptibility to the agent being evaluated compared to the WT (no phenotypic resistance) population. ECVs are calculated by taking into account the MIC distribution, the modal MIC of each distribution, and the inherent variability of the test (usually within one doubling dilution) and should encompass ≥97% of isolates. Because ECVs are based on MIC distributions generated using reference methods, ECVs based on other methods have the potential to be different. Nevertheless, it is possible to establish method-dependent ECVs using the same data requirements and techniques as for the SYO ECVs in the present study.

Data analysis.

Data were analyzed as previously described (13, 16–18). SYO MIC distributions for each species/echinocandin received from each center were listed in electronic spreadsheets. Pooled distributions were screened for (i) grossly skewed distributions (distributions that had a modal MIC [most frequent value] at the lowest or highest concentration tested and/or which were bimodal inside an apparent wild-type population), (ii) distribution size (data from ≥3 laboratories and the total pooled distribution had ≥100 isolates), (iii) unusual modal variation (global modes more than 2-fold dilutions from the other individual modes), and (iv) the percentage of isolates provided by each laboratory for each species/agent combination. The last criterion was important because if one of the laboratories included in the pool distribution for each species/agent provided ≥50% of the MIC data, the MIC distributions were weighted to reduce bias in the estimate. This was almost the case for anidulafungin versus C. albicans in the present study, where one laboratory contributed 48% of isolates. As a precaution, both weighted and unweighted data analyses were performed and produced the same ECV (0.12 μg/ml) (Table 3). Following the elimination of abnormal distributions, the resulting qualifying pooled distributions were used to calculate ECVs by the iterative statistical method (Tables 1 and 3), where the modeled population was based on fitting a normal distribution at the lower end of the MIC range, determining the mean and standard deviation of that normal distribution, and using those parameters to calculate the MIC that captured ≥97.5% of the modeled WT population (18). When the pooled MIC distribution was insufficiently symmetrical to use the statistical method effectively, a tentative ECV was estimated visually (C. albicans and micafungin).

TABLE 3.

Method-dependent SYO-ECVs of anidulafungin, caspofungin, and micafungin for eight species of Candida based on MICs from between 12 and 32 laboratories determined by the commercial Sensititre YeastOne broth microdilution method

Agent	Species	No. of isolates tested	MIC (μg/ml)		ECV (μg/ml)a
			Range	Modeb	SYO		CLSI, ≥97.5%
					≥95%	≥97.5%
Anidulafungin	C. albicans	3,118	0.016–2	0.03	0.12	0.12	0.12
	C. glabrata complex	2,520	0.016–≥8	0.03	0.12	0.12	0.12
	C. krusei	408	0.016–2	0.03	0.25	0.25	0.25
	C. parapsilosis complex	2,468	0.016–≥8	1	2	4	8
	C. tropicalis	652	0.016–8	0.12	0.25	0.5	0.12
	C. dubliniensis	131	0.016–4	0.12	0.25	0.25	0.12
	C. guilliermondii	101	0.016–≥8	1	4	4	8
	C. lusitaniae	185	0.016–1	0.12	0.25	0.25	1
Caspofungin	C. albicans	6,007	0.008–≥8	0.06	0.12	0.25	NAc
	C. glabrata complex	3,188	0.008–≥8	0.12	0.25	0.25	NA
	C. krusei	493	0.03–≥8	0.25	1	1	NA
	C. parapsilosis complex	3,136	0.008–≥8	0.5	2	2	NA
	C. tropicalis	1,016	0.008–≥8	0.06	0.25	0.25	NA
	C. dubliniensis	186	0.016–≥8	0.06	0.25	0.25	NA
	C. guilliermondii	118	0.06–≥8	0.5	2	2	NA
	C. lusitaniae	205	0.008–4	0.25	1	1	NA
Micafungin	C. albicans	4,862	<0.008–≥8	<0.008	0.06d	0.06d	0.03
	C. glabrata complex	2,763	0.008–≥8	0.016	0.03	0.03	0.03
	C. krusei	444	0.008–4	0.12	0.25	0.25	0.25
	C. parapsilosis complex	2,659	0.008–≥8	1	2	4	4
	C. tropicalis	983	0.008–2	0.03	0.06	0.06	0.06
	C. dubliniensis	123	0.008–≥8	0.03	0.06	0.12	0.12
	C. guilliermondii	119	0.016–4	0.5	2	2	2
	C. lusitaniae	191	0.008–0.5	0.06	0.12	0.12	0.5

^aCalculated SYO ECVs comprising ≥95% and ≥97.5% of the statistically modeled population and CLSI ECVs based on MICs determined by the CLSI M27-A3 broth dilution method as obtained in a prior study (13).

^bMost frequent MIC.

^cNA, not available.

^dSYO ECVs for micafungin and C. albicans calculated by visual inspection.

RESULTS AND DISCUSSION

BPs are essential for susceptibility testing to be clinically useful; they are reliable predictors of the clinical response to therapy. BPs have been established for the more prevalent Candida spp. for some antifungal agents by both CLSI and EUCAST (7, 11). More recently, ECVs have been statistically calculated for echinocandins and various species of Candida and Aspergillus using CLSI reference methods (13, 26). The main role of the ECV is to distinguish the two populations (WT and non-WT) that are present in the MIC distribution of a combination of species and agent (18). Therefore, ECVs can aid the clinician in identifying isolates that are potentially resistant or less likely to respond to therapy (non-WT isolates) when BPs are not available for the species/agent combination being evaluated. A previous study evaluated four methods for ECV calculation for echinocandins and four Candida spp. using SYO MICs (27). However, although isolates originated in numerous medical centers and some values are similar to those in the present study, only median tentative ECVs were listed; those values encompassed ≥95% of the isolates instead of the preferred ≥97.5% inclusion of the statistically modeled population. Therefore, for the first time, method-dependent ECVs for a commercial assay (SYO) were established according to the criteria set forth in the CLSI document under development for the establishment of ECVs. The rules will state that MICs must be available for ≥100 isolates for each species, the data must originate from at least three independent laboratories, and the ECV must be calculated using the iterative statistical method (16). In the present study, we gathered sufficient SYO MIC data from 38 laboratories to propose SYO ECVs for three echinocandins and eight species of Candida and to document MIC data for other, less common species. Although SYO MIC data were received from 41 laboratories, the distributions from some laboratories, depending on the antifungal agent and species, were omitted. These exclusions were due to either truncated (mode at the lowest concentration tested) or bimodal distributions or when MICs for the QC isolates were outside the recommended range (15). In addition, some laboratories provided data only for the more common species, such as C. albicans and C. parapsilosis complex. Most of the distributions for C. albicans and micafungin were truncated at the lower end, which precluded the statistical calculation of an ECV for this species/agent combination, because the mode of the pooled distribution was at the lowest concentration tested (Table 1). However, a tentative SYO ECV was provided based on visual inspection for this species/agent combination (Table 3).

The aggregated SYO MIC distributions of the three echinocandins and the eight most common species evaluated are listed in Table 1. Table 2 depicts the pooled distributions for less prevalent species for which >10 isolates were reported. Interestingly, the interlaboratory caspofungin modal variability (WT modes scattered six 2-fold dilutions) reported for the CLSI methodology was not observed in the present study by the SYO method (10). The SYO echinocandin modes from individual participant laboratories for the species evaluated were either the same as the overall mode or ±1 dilution, with only one exception. The mode of C. krusei and anidulafungin was two dilutions higher (0.12 μg/ml versus an overall mode of 0.03 μg/ml) in 2 of the 16 laboratories that provided data for this species (data not shown in Table 1). Overall, the lowest modes (≤0.016 and 0.03 μg/ml) were observed for C. albicans, C. glabrata complex, and C. krusei with anidulafungin and for C. albicans, C. glabrata complex, C. dubliniensis, and C. tropicalis with micafungin (Tables 1 and 3). As expected, the highest modes of these three echinocandins were for both C. parapsilosis complex and C. guilliermondii (0.5 to 1 μg/ml). Our results reflect those of previous studies where echinocandin MIC data from multiple laboratories were evaluated for the four more common species regarding both similar modes (13, 14) and the lack of caspofungin modal variability (14). Anidulafungin and micafungin modes obtained by the CLSI broth microdilution method for the eight species for which sufficient data allowed ECV definition were also similar (13). To our knowledge, pooled SYO MIC data are not available for the less common Candida spp. depicted in Table 2. Overall, the lowest modes were for C. pelliculosa (0.03 μg/ml) versus anidulafungin and micafungin and the highest for C. orthopsilosis (0.5 to 1 μg/ml) with the three agents, as well as C. lipolytica and C. rugosa (0.5 μg/ml) with caspofungin, but the distributions were small.

Table 3 summarizes the proposed anidulafungin, caspofungin, and micafungin SYO ECVs using both ≥95% and ≥97.5% of the modeled MIC wild-type distributions. MIC ranges and modal MICs for the eight species of Candida spp. with the required number of isolates/laboratories for SYO ECV calculation are also depicted in Table 3. The CLSI recommends ≥97.5% as the preferred percentage for establishing ECVs (versus the ≥95% values) in the future guidance document under development for this purpose (16). Using the lower percentage risks classifying some WT isolates as non-WT isolates (major error), while the using the higher percentage risks classifying some isolates with acquired resistance mechanisms as WT (very major error). Although both ≥97.5% and ≥95% values are listed, the discussion here focuses on the higher values for the aggregated distributions of the eight Candida spp. with sufficient data. The ECV of 0.06 μg/ml for C. albicans and micafungin had to be estimated visually and is therefore a tentative value; the excessive amount of truncated MIC distributions (modes at the lowest concentration tested) precluded statistical analysis for this species/agent combination. In general, ECVs for each species of Candida and antifungal agent were within 1 or 2 2-fold dilutions of the mode values. The lack of reference caspofungin ECVs due to modal variability (10, 12) and the listing of several ECVs (according to the method used for their calculation) in a previous study (27) precluded meaningful comparisons. As previously reported for CLSI anidulafungin and micafungin ECVs (13), our SYO anidulafungin ECVs were mostly 1 to 2 2-fold dilutions higher than those for micafungin. Examination of both ≥97.5% CLSI and SYO ECVs for both anidulafungin and micafungin for the eight Candida spp. indicated that values were the same or ±1 dilution, with three exceptions (13). A higher SYO ECV was calculated for anidulafungin and C. tropicalis (0.12 μg/ml and 0.5 μg/ml, respectively) and lower SYO ECVs for C. lusitaniae versus both anidulafungin (0.25 μg/ml and 1 μg/ml, respectively) and micafungin (0.12 μg/ml and 0.5 μg/ml, respectively).

Table 4 presents the application of the echinocandin SYO ECVs, encompassing ≥97.5% of the statistically modeled population, and the ability of these values to identify strains of Candida isolates with intrinsic or acquired mechanisms of resistance (non-WT). The total of 81 molecularly defined mutants (i.e., isolates with fks1 [all species] or fks2 [C. glabrata only] gene mutations) included 41 C. albicans, 26 C. glabrata, 5 C. krusei, 8 C. tropicalis, and 1 C. dubliniensis isolates (8, 19–24); SYO MICs for the WT isolates are also depicted in Table 4. SYO ECVs of anidulafungin of 0.12 μg/ml for C. albicans and C. glabrata complex, 0.25 μg/ml for C. krusei and C. dubliniensis, and 0.5 μg/ml for C. tropicalis correctly classified 72 of the 81 mutants as non-WT (MICs were greater than the ECV). In the same manner, SYO micafungin ECVs of 0.03 μg/ml for C. glabrata complex, 0.06 μg/ml for C. albicans (tentative value) and C. tropicalis, 0.12 μg/ml for C. dubliniensis, and 0.25 μg/ml for C. krusei correctly identified 76 of the 81 mutants as non-WT. These results are similar to those obtained for anidulafungin CLSI ECVs (88.9 versus 92.2%, respectively) but lower for micafungin (93.8 versus 100%) (13). Therefore, both echinocandin SYO and CLSI ECVs appear to have similar predictor values in identifying the mutants. Caspofungin ECVs of 0.25 μg/ml for C. albicans, C. glabrata complex, C. tropicalis, and C. dubliniensis and the ECV of 1 μg/ml for C. krusei classified as non-WT 74 (91.4%) of the 81 mutants, results similar to those for micafungin ECVs in the present study. The presence of mutations in isolates for which echinocandin MICs are low has been documented in various studies where the relationship between resistance mechanisms, MICs obtained by different methodologies (the SYO method included), and response to therapy has been evaluated, especially for C. glabrata (e.g., values of ≤0.12 μg/ml, ≤0.25 μg/ml, and ≤0.03 μg/ml anidulafungin, caspofungin and micafungin, respectively, for mutant isolates) (9, 28–30). Although not all fks mutations confer phenotypic resistance, the same mutation was present in two isolates of C. albicans (FSK1 S645F) and three isolates of C. glabrata (FSK2 F659L) for which low echinocandin SYO MICs were obtained (Table 4). Therefore, since testing the susceptibilities of Candida spp. to caspofungin by reference methodologies is not recommended yet standardized (7, 10, 12), the SYO method offers a solution to this problem. Due to the lack of reliable caspofungin species-specific BPs or CLSI ECVs for Candida spp., our SYO ECVs should aid laboratory personnel as well as physicians in identifying those non-WT isolates having presumptive acquired resistance mechanisms.

TABLE 4.

Application of SYO ECVs of anidulafungin, caspofungin, and micafungin to identify Candida isolates harboring fks1 and fks2 gene mutations (non-WT strains)^a

Agent	Species	No. of mutants identified/no. of mutants (%)b	SYO ECV (μg/ml)c	MIC for wild type (μg/ml)d
Anidulafungin	C. albicans	36/41 (87.8)	0.12	0.016–0.03
	C. glabrata	24/26 (92)	0.12	0.016–0.12
	C. krusei	5/5 (ND)	0.25	NA
	C. tropicalis	6/8 (87.5)	0.5	0.03–0.06
	C. dubliniensis	1/1 (ND)	0.25	NA
	Total	72/81 (88.9)
Caspofungin	C. albicans	38/41 (92.7)	0.25	0.06–0.12
	C. glabrata	24/26 (92)	0.25	0.06–0.25
	C. krusei	4/5 (ND)	1	NA
	C. tropicalis	7/8 (87.5)	0.25	0.016–0.25
	C. dubliniensis	1/1 (ND)	0.25	NA
	Total	74/81 (91.4)
Micafungin	C. albicans	38/41 (92.7)	0.06e	0.008–0.03
	C. glabrata	24/26 (92)	0.03	0.016–0.03
	C. krusei	5/5 (ND)	0.25	NA
	C. tropicalis	8/8 (100)	0.06	0.03–0.06
	C. dubliniensis	1/1 (ND)	0.12	NA
	Total	76/81 (93.8)

^aMutation data are from references 8 and 19, to ,24. ND, not determined due to small values; NA, not available.

^bGene mutations present in the isolates misclassified as wild type were as follows: C. albicans, FKS1 L644L/stop, F641L, R136IR/H, and S645F (2 isolates); C. glabrata, FKS2 F659L (3 isolates), F659S, and S663P; C. krusei, FKS1 H675H/Q; and C. tropicalis, FKS1 F641L, and F641S.

^cCalculated SYO ECVs comprising ≥97.5% of the statistically modeled population.

^dSYO MICs for wild-type isolates (2 C. albicans, 6 C. glabrata, and 6 C. tropicalis).

^eSYO ECV for micafungin and C. albicans calculated by visual inspection.

In conclusion, we propose species-specific SYO ECVs for eight different species of Candida versus anidulafungin, caspofungin, and micafungin based on MIC data originating from nine different countries. Although SYO ECVs that encompass ≥97.5% of the statistically modeled population as recommended by the CLSI are provided, we also have calculated ECVs encompassing ≥95%; some of the latter values are lower. These ECVs would differentiate WT from non-WT strains of C. albicans, C. dubliniensis, C. glabrata complex, C. guilliermondii, C. krusei, C. lusitaniae, C. parapsilosis complex, and C. tropicalis and the three echinocandins when susceptibility testing is performed by using the SYO colorimetric method. Echinocandin ECVs could also help to monitor for the emergence of echinocandin resistance among target species of Candida. The potential clinical usefulness of these species-specific SYO ECVs was demonstrated by their ability to classify 88.9 to 93.8% of 81 well-characterized fks mutant strains of five Candida spp. This is especially important in the case of caspofungin, since at this time MICs obtained by reference methodologies for Candida spp. are not reliable. However, further evaluation of the proposed SYO ECVs is needed for those species for which only a few mutant isolates were available.