Background

Candida albicans once accounted for most serious nosocomial candidal infections. Over the last decade Candida infections, specifically candidemia, due to C. albicans have declined [1]. Other non-albicans Candida species (NAC) are now responsible for about half of candidemias and other deep Candida infections [2–4]. Most NAC infections are caused by C. glabrata, C. parapsilosis, or C. tropicalis, although in total over a dozen NAC have been reported to cause candidemia and other invasive infections [5–8]. In addition to deoxycholate and lipid formulations of amphotericin B, clinicians now must chose from azoles with differing spectrums of activity (fluconazole, itraconazole, and voriconazole) and echinocandin antifungal agents. In vitro testing has revealed that there are clear differences among the various NAC in their susceptibility to specific drugs. Rapid, reliable identification to species is now needed more than ever for clinicians to make treatment choices.

Identification of yeast pathogens by traditional methods requires several days and specific mycological media. Chromogenic media contain chromogenic substrates which react with enzymes secreted by the target microorganisms to yield colonies of varying colors. CHROMagar Candida (CaC) is one such medium that, per its manufacturer (CHROMagar Microbiology, Paris, France, http://www.chromagar.com), can identify three candidal yeasts, C. albicans (green colonies), C. tropicalis (steel blue colonies), and C. krusei (fuzzy, rose colored colonies) after 48 hours of incubation at 30–37°C. Independent groups have reported success with CaC in differentiating C. dubliniensis from C. albicans [9, 10]. Investigators have also reported the ability to distinguish C. glabrata from other species [11–14], although this has been contested by other authors [15–17].

We studied the appearance of 180 isolates of 12 Candida species on CaC to determine whether these more rare species could be reproducibly distinguished from each other with this medium.

Methods

Clinical yeast isolates were employed for all testing, predominately from our facility. Additional isolates were provided by the Fungus Testing Laboratory, University of Texas Health Science Center at San Antonio, Texas. Yeasts tested included C. dubliniensis (17 isolates), C. famata (11 isolates), C. firmetaria (12 isolates), C. glabrata (38 isolates), C. guilliermondii (10 isolates), C. inconspicua (6 isolates), C. kefyr (9 isolates), C. lipolytica (10 isolates), C. lusitaniae (15 isolates), C. norvegensis (3 isolates), C. parapsilosis (34 isolates), and C. rugosa (15 isolates). Five isolates each of C. albicans, C. krusei, and C. tropicalis were used as controls. All isolates were identified by standard clinical practices as previously described [18].

Inocula from clinical isolates stored at -70°C were transferred into yeast extract peptone (YEP) broth (Bacto Peptone, BD, Sparks, MD) and incubated at 30°C in a rocking incubator for up to 96 hours. Recovered yeasts were then streaked for isolation onto Sabouraud dextrose agar (SDA) plates (BD, Sparks, MD) and incubated at 30°C for 24–48 hours and subcultured a second time prior to inoculation in duplicate onto commercially prepared CHROMagar Candida (CaC) plates (BD, Sparks, MD). Inoculated CaC plates were incubated in parallel at 30°C and 37°C. Two independent readers, blinded to species inoculated, observed each set of plates for color and colony morphology daily for 7 days. Following the initial blinded study, select isolates were repeated in an unblinded fashion in parallel with the control isolates for direct comparison.

Results

Individual isolates produced the same colors at both incubation temperatures. In general, these colors appeared more intense at 37°C and intensified daily at both temperatures, peaking at 72 hours. Most of the NAC (C. famata, C. firmetaria, C. guilliermondii, C. inconspicua, C. kefyr, C. lipolytica, C. lusitaniae, C. norvegensis, and C. parapsilosis) tested produced variable shades of ivory, pink, and lavender (Table 1). C. parapsilosis isolates were the most varied of all species tested, in both color and colony morphology. C. kefyr commonly produced large colonies with centers that deepened to a brownish color with duration of incubation. C. firmetaria and C. inconspicua isolates produced large flat colonies. Most of the C. firmetaria and all of the C. inconspicua were pink with pale borders and rough in texture, indistinguishable from the color and morphology of C. krusei. C. rugosa most commonly (eleven of 15 isolates) produced light blue-green colonies with pale borders, similar in morphology (large, flat) to C. krusei. The minority of isolates grew as small to medium pink colonies not distinguishable from many other species.

Table 1 The appearance of Candida species grown on CHROMagar Candida.
Full size table

C. glabrata was readily identifiable using CaC, producing small, convex, dark pink to violet colonies. Peak color intensity was noted at 72 hours when incubated at 30°C and at 48 hours when incubated at 37°C. C. glabrata isolates typically produced colonies with thin pale borders and a deep violet pigment that diffused into the medium. Unfortunately, neither of these features were universal. Our readers could not distinguish the medium to dark green colonies produced by C. dubliniensis from those of C. albicans at either temperature or any duration, except when directly compared to each other on repeat testing.

Discussion

Amphotericin B, fluconazole, and caspofungin are currently the three antifungal agents most commonly used to treat candidemia and invasive candidiasis. This practice is supported by FDA approval and the most recent Infectious Diseases Society of America guidelines [19]. While all these agents have been used effectively in clinical studies, these studies included mostly patients with C. albicans infections; thus their efficacy against NAC is not necessarily universal or well known. In vitro study has shown decreased susceptibility to amphotericin B in isolates of C. famata, C. guilliermondii, C. inconspicua, C. kefyr, C. krusei, C. lusitaniae, and C. rugosa and to fluconazole in isolates of C. famata, C. firmetaria, C. guilliermondii, C. inconspicua, C. krusei, and C. lusitaniae [7, 8]. The MICs of caspofungin have been demonstrated to be higher in C. famata, C. guilliermondii, and C. parapsilosis isolates [20]. The number of antifungal agents available continues to increase in the setting of a shift of candidal infections to those caused by non-albicans species. Because of this, identification to species and increased use of susceptibility testing has become necessary to appropriately select which agent to use [21].

CHROMagar Candida is a chromogenic medium that is advertised as able to identify C. albicans, C. krusei, and C. tropicalis. With the increasing incidence of human disease being produced by the less common Candida species, we were interested in testing the performance of this medium with the less common agents of candidiasis. We found, as previously reported, that most of these rarer Candida species and C. parapsilosis produced typical convex, creamy yeast colonies in shades of pink, lavender, and less commonly ivory, not distinguishable from each other (and often not consistent between isolates of the same species). C. parapsilosis, one of the most commonly recovered NAC, produced the widest range of colors and morphologies, making it impossible to identify using this medium.

Many of the rare NAC produced morphology and colors similar to those seen with C. krusei on this medium. We found individual isolates of C. lipolytica and C. norvegensis, many isolates of C. firmetaria, and all tested C. inconspicua were indistinguishable from C. krusei. Some that were distinguishable, produced similar color as that of C. krusei, with the pale border, but with waxy colonies. Most of these species share reduced susceptibility to fluconazole that is common with C. krusei.

We confirmed our previous observation that C. rugosa appears to typically produce a readily identifiable and unique color/colony type on CaC [12, 22]. Eleven of 15 isolates of this organism produced the same light blue-green color and a colonial morphology similar to that of C. krusei. C. rugosa has been shown in clinical reports and by in vitro testing to be less susceptible to amphotericin B [7, 8, 23]. Rapid identification of this NAC is of great importance to allow provision of appropriate therapy to patients.

As previously reported by our group and others, this study did find C. glabrata to be readily distinguishable from other species that produced pink to lavender colonies on CaC. With the use of larger numbers of isolates, we again noted that this species produced smaller colonies starting out as dark pink hues, becoming dark violet with time, commonly with a small diffusion of dark violet pigment into the surrounding agar and a thin pale border [12, 22]. This was most apparent after prolonged incubation (72 hours) at the lower temperature suggested by the manufacturer (30°C), or after 48 hours when incubated at 37°C.

Our readers could not distinguish C. dubliniensis from C. albicans when observed in a blinded fashion. The isolates we tested were somewhat darker, especially at the higher incubation temperature, but this was only obvious when CaC culture was performed in parallel to C. albicans controls. The report by Kirkpatrick et al. [9] that made the initial observation that C. dubliniensis produce darker green colonies with CaC did this with primary isolation of clinical materials onto the medium. Our isolates were not from primary samples and thus may have been affected by storage conditions and repeated subculturing. Odds and Davidson reported that they could differentiate C. dubliniensis from C. albicans using stored isolates, but this was best shown after 72 hours of incubation at 37°C [10].

Conclusion

In our hands CHROMagar Candida shows good potential as a screening medium to rapidly identify potentially azole-resistant as well as amphotericin B-resistant species, including C. krusei, C. glabrata. C. rugosa, and C. inconspicua. Identification of C. dubliniensis appears to have some limitations based on our study, but CaC appears (from other studies) to also hold this potential when used as a primary medium. We would suggest inclusion of a C. albicans control plate if this medium is used as a primary fungal recovery medium to help increase the presumptive identification of C. dubliniensis. Combined with confirmation of speciation by standard methods and knowledge of the local antifungal susceptibility patterns of these species, CaC may be a useful adjunctive medium for use in the clinical laboratory.