Strengths and Limitations of different Chromogenic Media for the Identification of Candida Species

The treatment of invasive candidiasis and other Candida infections with the appropriate antifungal agent is assisted by the identification of Candida isolates to the species level. Rapid and accurate methods of differentiation are therefore imperat ive if t reatment is to be effective, particularly in HIV-positive patients and in pregnant mothers where intervention may be necessary to reduce the risk for preterm delivery. The time used for isolation, identificat ion and detection of mixed cultures may be reduced with the help of available chromogenic media. In this study, five commercial chromogenic media were evaluated for the differentiation of Candida species. Six type-strains of Candida species were streaked onto each of five different chromogenic media and incubated for up to 4 days at the different temperatures recommended by the manufacturers. This comparative evaluation demonstrated the strengths and weaknesses of each medium employed and found CHROMagarTM Candida and Chromogenic Candida Agar to be the most effective for distinguishing between different Candida species.


Introduction
There has been a significant increase in the number of Candida resistant cases in hospital patients in the last 20 years. Predisposing factors include particularly prolonged and increased use of antifungal agents [1] and patients with compro mised immune systems, such as HIV-positive patients [2] and pregnant mothers with asymptomat ic vaginal candidiasis who run the risk of preterm delivery [3]. Amongst the species most frequently isolated are Candida albicans followed by Candida glabrata, Candida tropicalis and Candida krusei [4].
In health, Candida albicans is a harmless commensal fungus, while, in immunocompro mised patients, it may cause superficial o r even life-threatening systemic infections [5]. It is not entirely understood how the mechanis ms of change from a non-pathogenic to a pathogenic phenotype occurs. Knowledge of the metabolic activity of Candida albicans rema ins limited even though a great deal of research has been done on aspects of its pathogenicity [5].
Candida dubli nien sis is a fairly recent ly described species of Candida with similar characteristics to that of Candida albicans. It is clin ically important to co mpare the pathogenesis and management o f infect ion by a newly discovered species, with infection caused by other me mbers of the same genus [6]. Candida albicans and Candida dubliniensis have the same morphological and physiological characteristics due to the close association in their phylogenetics, e.g. germ-tube and chlamydospore formation [6]. This has caused a problem in d ifferentiating between the two species, with the result that Candida dubliniensis strains have been, and will continue to be, identified in the clin ical laboratory as Candida albicans [6]. To ma ke a precise differentiation between the two species requires PCR-based tests, but due to the high quantities of throughput samples at diagnostic laboratories, this is not feasible and thus PCR-based tests are mostly used in research laboratories [7]. Looking at the phenotypic characteristics is much mo re inexpensive than that of the genotypic characteristics, and scientists have therefore demonstrated the use of selective and differential media for the presumptive identificat ion of Candida species with good sensitivity and specificity [8], thereby reducing the time used for isolation, identification and detection in mixed cultures [9].
The purpose of this study was to perform a co mparat ive evaluation of five different chro mogenic media in order to establish which would y ield the most reliab le d ifferentiation of frequently isolated Candida species namely, Candida albicans, Candida dubliniensis, Candida tropicalis, Candida krusei and Candida glabrata.

Type-strains of Candida Used
A total of six type-strains of Candida species were used for the evaluation of the five chro mogenic media. Of these type-strains, C. albicans (ATCC 90028), C. tropicalis (ATCC 950), C. krusei (ATCC 2159), C. glabrata (ATCC 26512) were obtained fro m the A merican Type Cu lture Collection (ATCC, Manassas, VA, USA.) and C. albicans (NCPF 3281) and C. dubliniensis (NCPF 3949a) fro m the National Co llect ion of Pathogenic Fungi (NCPF, Bristol, United Kingdom). These type-strains were stored in fro zen stocks in cryovials at -70 ℃ and cultured twice on Sabouraud's dextrose agar (Oxoid , CM 0041) for 24 hours prior to the inoculation of the chro mogenic med ia.

Inoculation of Chromogenic Me di a
Chro mogenic med ia used, included commercially prepared CandiSelect™4 Agar (Bio-Rad, 63746) wh ile, Chro mogenic Candida Agar (Oxoid, CM1002A), Bismuth Sulphite Glucose Glycine Yeast agar (BiGGY Agar) (Oxoid, CM0589B) also known as Nickerson's mediu m [10], modified Candida Ident Agar, (Fluka, 94382) and CHROMagar™ Candida (CHROMagar, CA 220) were purchased in a dehydrated form and prepared according to the manufacturers' instructions. All plates were left to reach room temperature prior to inoculation if previously stored at -4℃. Type-strains of Candida species were inoculated onto the different chro mogenic media and each incubated for up to 4 days at the different temperatures recommended by the manufacturers. Th is was done in triplicate. CandiSelect ™4 Agar and CHROMagar™ Candida were incubated at 37℃, modified Candida Ident Agar, and Chro mogenic Candida Agar were incubated at 30 ℃ , and BiGGY Agar was incubated at 28-30℃. The plates were checked after 24, 48, 72 and 96 hrs for growth to determine when (according to the manufacturers' c laims) the expected colour, morphology or texture of the colonies appeared, and whether prolonged incubation would affect the results.

Statistical Analysis
Because of the small sa mple size no meaningfu l statistical analyses could be performed.

Results
All the type-strains grew on the five different chromogenic med ia. So me type-strains were mo re distinguishable than others. The appropriate colour, textu re and morphology of the colonies were observed after each 24-hour period for a total of 96 hours and compared with the recommended time period of the manufacturers. So me chromogenic media characterized the different type-strains by colour only while others characterized them by colour, texture and morphology.
Both C. albicans type-strains (ATCC 90028 and NCPF 3281) appeared as predicted on CHROMagar™ Candida, modified Candida Ident, and Chro mogenic Candida Agar (Table 1). They appeared as pink colonies after 24 hours, which darkened to purple after incubation for 48 hours on CandiSelect ™4 Agar. On BiGGY Agar, the predicted colour reactions for C. albicans were expressed, while the expected mycelial fringe was not observed even after prolonged incubation of 96 hours. (Table 1) Colonial mo rphology of C.dubliniensis differed fro m the predicted patterns for all 5 of the chro mogenic med ia used ( Table 2). A lthough the guidelines for CHROMagar™ Candida, modified Candida Ident Agar and BIGGY Agars predicted that C.dubliniensis could not be distinguished, results on the CHROMagar™ Candida revealed that C. albicans and C. dubliniensis could clearly be d istinguished with C. albicans colonies yielding a green-turquoise colour while C. dubliniensis appeared plain green after 48 hours incubation ( Table 2). After a longer incubation period (96 hours), no change was observed in C. albicans while colonies of C. dubliniensis formed a darker centre, a characteristic clearly distinguishing it fro m C. albicans (Fig.  1a,b).
Chro mogenic Candida Agar guidelines predicted a green colour, but we observed translucent light-blue colonies after 24 hours which intensified to dark blue on prolonged incubation of 96 hours (Fig.1c,d). Prolonged incubation was also required for CandiSelect™4 Agar since the pink-purp le colonies predicted after 24 hours only appeared after 72 hours of incubation (Table 2).    (Table 3), while Candiselect ™4 Agar yielded the predicted pale turquoise colonies after 24 hours, which darkened to deep turquoise centred colonies with white peripheries after 48 hours. Chro mogenic Candida Agar produced beige-cream to light brown colonies. However, this did not distinguish them fro m other Candida species but when incubated for longer than 72 hours, the colonies started to turn pink.
All of the 5 chro mogenic media y ielded the predicted results for C. krusei (Table 4). Although the guidelines mention silver b rown-black, we concede that the reflection of the light in the dark bro wn colonies could have been interpreted by us as gold rather than silver.
The CHROMagar™ Candida and modified Candida Ident Agar, guidelines predict a metallic blue colony for C.tropicalis, but we observed dark-purple blue colonies on the CHROMagar™ Candida (Table 5) and light lilac colonies on modified Candida Ident Agar after 24 hours, which intensified to blue after 48 hours. Neither of the agars grew colonies with a metallic sheen. Gro wth on CandiSelect ™4 Agar appeared to match the overall morphology as described in the guidelines, but the colonies appeared blue and not turquoise in colour. Likewise, the colonies appeared to be similar to the BiGGY Agar guidelines, but no mycelial fringe was evident, nor did the med ia blacken after 72 hours.
Following the pilot study using only the type-strains, clin ical strains fro m our laboratory collection, prev iously identified as C. albicans, C. dubliniensis, C. krusei, C. glabrata and C. tropicalis were also compared for consistency in the evaluation of the chro mogenic media. Colony colour and mo rphology observations from the different clin ical strains showed the same results as the typestrains for all chro mogenic agars.

Discussion
This study evaluated CHROMagar™ Candida, Candida Ident Agar (modified ), Chro mogenic Candida Agar, CandiSelect ™4 Agar and BiGGY Agar for their efficacy in the presumptive identificat ion and differentiat ion of Candida species. An appropriate primary cu lture med iu m that assists in the recovery and differentiation of colonies which are phenotypically similar is a vital requirement for the laboratory detection of mixed fungal clinical specimens. Traditional methods for identification of yeast pathogens involves several days and specific mycology media wh ile chromogenic med ia contains chromogenic substrates which react with enzy mes secreted by the organisms to give colour reactions for d ifferent species [9] thus complementing traditional methods of identification [11]. CHROMagar™ Candida is the most well-known and widely used chromogenic mediu m for the identificat ion of different Candida species and is the most expensive of the five chromogenic media . Results fro m mixed cultures are reported to provide results 24 to 48 hours sooner than standard isolation and identification methods. It contains a variety of substrates which interact with the enzymes secreted by the yeast species and has been reported to selectively isolate and identify Candida species with a high degree of accuracy [12] sensitivity and specificity [13].
As in our study, previous studies reported green colonies for C. albicans [12], [14] dark b lue colonies for C. tropicalis and pink co lonies with a downy appearance for C.krusei [8].
Although not clearly distinguishable after 48 hours, prolonged incubation (96 hours) proved useful for differentiating C.albicans from C.dubliniensis. Modified Candida Ident Agar, is a new chromogenic mediu m on which, we assume, not much research has been done. In this study, modified Candida Ident Agar and CHROMagar TM differentiated between the different Candida species by colour only. C. krusei however, could be differentiated by both colour and texture on both media. With the exception of C. glabrata, a more accurate colour exp ression of the other three species occurred after 48 hours of incubation, which suggested that the colours and texture description of the specific species of Candida presented by Candida Ident Agar would have been more accurate following an incubation of 48 hours rather than 24 hours. These results confirm that this mediu m does not reflect the appropriate results suggested by the manufacturer and therefore is not as effective in the differentiation of Candida species.
Chro mogenic Candida Agar (Oxo id) has been re-named "Oxo id Brilliance Candida Agar" but in this study, we refer to it as "Chromogenic Candida Agar". It is a new commercial ready-to-use chromogenic med iu m, contains chromogenic substrates which react with the different enzy mes of species of Candida, such as hexosaminidase and alkaline phosphatase resulting in the expression of a specific colour in the colony. The different colours appear as a result of different enzy mes produced by the different species [15]. C. albicans, C. dubliniensis and C. tropicalis produce the enzy me hexosaminidase which results in the colonies being green, but C. tropicalis yields dark blue colonies due to other metabolic reactions causing a drop in p H [15]. C. krusei yielded brown or pink colonies because it produces alkaline phosphatase and due to a combination of natural pigmentation and some alkaline phosphatase activity, C. glabrata yielded a variety of natural colour, such as beige, brown and yellow.
CandiSelect ™4 Agar (Bio-Rad) contains two chromogenic substrates which interact with hexosamin idase and phosphatase produced by the different Candida species [4], while a co mbination of antibiot ics such as chloramphenicol and gentamicin may suppress bacterial growth. In this study, C. albicans, C. krusei and C. glabrata yielded results described by the manufacturer while the other type-strains did not, thus questioning the reliability of this med iu m. Each of the different species of Candida requires different incubation periods on this medium. Similar results have been reported [4] for C. krusei, while identification o f C. tropicalis and C. glabrata were regarded as presumptive only.
In this study, BiGGY Agar was not able to distinguish the different Candida species due to the fact that all the type-strains were in the same colour range and that the distinctive characteristics such as the mycelial fringe scarcely occurred and when it did occur, it was never at the recommended incubation period.
We have just touched on the differentiat ion between Candida albicans and Candida dubliniensis, since, in addition to looking at the other co mmonly isolated Candida species, we were also interested in establishing whether adj usting incubation times of chro mogenic med ia could adequately differentiate between Candida albicans and Candida dubliniensis. We believe that we have achieved this.

Conclusions
The expression of antifungal susceptibility among different Candida species and the misidentificat ion of C. dubliniensis as C. albicans highlights the potential clinical importance of accurate species differentiation. The use of chromogenic med ia for the rap id and effective identification of Candida species has gained popularity within the clinical laboratory but presents with limitations in that inaccuracies often occur between the reactions described by the manufacturer and the actual results obtained in the laboratory. Differences in colonial morphology may occur as a result of differences in the laboratory conditions under-which the experiments are conducted e.g. the water used for med ia preparation may be of a different purity, thus affecting the substrate in the mediu m and thereby producing a different colour expression for specific species. Differences in co lour and reflection perceptions by different examiners should also be taken into account. By emp loying several chromogenic med ia and optimising the incubation periods for each species, sometimes deviating fro m the recommendations of the manufacturers, we were ab le to establish which med ia produced the most reliab le and consistent results and thus accurately differentiate the Candida species commonly infecting HIV-positive indiv iduals and pregnant Candida-infected mothers.
This comparative evaluation proved that CHROMagar™ Candida and Chro mogenic Candida Agar were the most effective of the chromogenic media evaluated and both yielded the expected colour colonies at the expected time period of incubation as suggested by the manufacturer. Candida Ident Agar (modified) and CandiSelect ™4 Agar only yielded results typical of three of the type-strains as suggested by the manufacturer, while BiGGY Agar y ielded all of the type-strains in one colour range and none of the differentiating morphological characteristics predicted were ever observed. In order to eliminate inaccuracies in the presumptive identificat ion of C. dubliniensis, we strongly support the use of CHROMagar™ Candida since this med iu m most clearly demonstrated the difference between Candida albicans and Candida dubliniensis thus reducing error in the identification of the two species.