INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Fungal isolates have emerged as important pathogens in the last two decades (14). The increase in the incidence of fungal infections has been associated with the increase in the number of patients at risk, i.e., patients with impaired cell-mediated immunity, such as human immunodeficiency virus-infected or cancer patients (10, 23). Candida albicans is by far the most frequent agent responsible for fungal infections; however, the emergence of non-C. albicans species, such as Candida parapsilosis, Candida krusei, and Candida tropicalis, has also been observed (3, 14, 23, 24).

Recently Sullivan and colleagues described a new species of the genus Candida, Candida dubliniensis, which is now recognized as a minor constituent of normal human oral microbial flora (4, 20, 22). Although most of the C. dubliniensis isolates have been recovered from the oral cavities of human immunodeficiency virus-infected patients, this fungal organism has also been isolated from specimens from different body sites, including lungs, vagina, blood, and feces (9, 13, 22).

To have a better understanding about the clinical significance and epidemiological role displayed by C. dubliniensis in human infections, it is essential that microbiology laboratories can identify this species rapidly and accurately in clinical specimens, utilizing reliable tests. This task is difficult due to the high degree of phenotypic and genotypic similarity between C. dubliniensis and C. albicans isolates (22). Frequently this great similarity has contributed to the misidentification of C. dubliniensis as C. albicans (4, 9, 21). Indeed, retrospective analysis of Candida stock collections has revealed that the oldest known isolates of C. dubliniensis were recovered in the 1950s. One isolate recovered in the United Kingdom in 1957 was originally misidentified as Candida stellatoidea (20, 21), and a second isolate recovered in Holland in 1952 was originally misidentified as C. albicans (9).

Although C. dubliniensis and C. albicans isolates are both susceptible to azoles, fluconazole resistance has been observed in clinical isolates of C. dubliniensis from AIDS patients with prior exposure to fluconazole (11, 12, 16). In addition, stable fluconazole resistance can be readily induced in C. dubliniensis isolates following direct exposure to the antifungal drug in vitro (11, 12). Consequently, these findings may have implications for antifungal therapy and indicate another important reason for distinction between these two species.

Several tests based on phenotypic characteristics have been utilized to distinguish C. dubliniensis from C. albicans. The tests evaluate -glucosidase activity, carbohydrate assimilation, colonial coloration on CHROMagar Candida agar medium, fluorescence on methyl blue-Sabouraud agar, and growth at 45°C (3-5, 7, 17-19). Polyclonal antibodies against C. dubliniensis have also been used for identification of this species (1). However, many of these tests have subsequently proved to be unreliable for identification and differentiation of C. dubliniensis (7, 17). Only molecular methods such as DNA fingerprinting and hybridization analysis with molecular probes have performed well enough to discriminate between these two species (4, 6, 15). Recently Joly and colleagues (6) have described a C. dubliniensis-specific DNA fingerprint probe, Cd25, that should prove to be useful both as a means of identifying this species and for epidemiological studies. However, these techniques are not suitable for application to large numbers of clinical isolates in routine diagnostic laboratories.

It has been previously suggested that isolates of C. dubliniensis could be differentiated from C. albicans by their inability to assimilate xylose (XYL) and -methyl-D-glucoside (MDG) (18). The principal objective of this study was to evaluate the performance of the XYL and MDG tests contained in the Vitek and the API 20C AUX systems to distinguish C. dubliniensis from C. albicans isolates. The ability to grow at 45°C was also studied.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Specimen collection and identification tests. A collection of 66 isolates of C. dubliniensis from the culture collection of The University of Dublin, Dublin, Ireland, and 100 clinical isolates of C. albicans from the bank collection of the Molecular Epidemiology and Fungus Testing Laboratory, University of Iowa, were evaluated in this study. The C. dubliniensis isolates were previously identified based on their ability to produce germ tubes and chlamydospores, by DNA fingerprinting analysis, by karyotyping, by indirect immunofluorescence with C. dubliniensis blastospore-specific polyvalent antiserum, and by hybridization with the C. dubliniensis-specific probe Cd25 (1, 3, 6, 16). The identifications of the C. albicans isolates were previously confirmed by hybridization with the C. albicans-specific probe Ca3 (15).

Culture. The yeast isolates were cultured on CHROMagar Candida plates (Hardy Diagnostics, Santa Maria, Calif.) prior to testing to ensure viability and purity. Yeast inoculum suspensions were prepared from 48-h cultures grown on CHROMagar Candida at 30°C. Briefly, yeast cells were suspended in 2 ml of a saline solution (0.45%) to achieve a turbidity of a no. 2 McFarland standard. The same inoculum suspension of each isolate was used by both systems.

API 20C AUX system. According to the instructions of the manufacturer (bioMérieux, Hazelwood, Mo.), 100 µl of inoculum suspension was transferred to the API basal medium ampoules. The API trays were inoculated and incubated for 72 h at 30°C. Cupules showing turbidity significantly greater than that of the negative control cupule were considered positive.

Vitek system. YBC cards were inoculated by the Vitek system (bioMérieux) with the inoculum suspension required. The YBC cards were incubated at 30°C for 24 h and submitted for reading. As detailed in the manufacturer's instructions, some YBC cards were incubated for an additional period of 24 h, giving a total incubation period of 48 h.

Reproducibility tests. Ten C. dubliniensis isolates and 10 C. albicans isolates were randomly chosen for reproducibility tests. Each isolate was tested three times after its initial result, on three different days, with different lots of API trays and Vitek cards.

Growth at 45°C. A small portion of a single colony of each isolate was removed from the CHROMagar Candida plate, streaked for isolation over the surface of a potato dextrose agar plate (Remel, Lenexa, Kans.), and incubated at 45°C. The growth was observed after 48 h. If the growth of the colonies extended into the last three quadrants of the plate, it was considered good growth. However, if the growth area was observed only in the first quadrant, it was considered poor growth.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The ability to grow at 45°C and the results for the XYL and MDG tests contained in the API 20C AUX and Vitek systems are shown in Table 1. None of the 66 C. dubliniensis isolates tested was able to grow at 45°C. Among the 100 C. albicans strains, 23 were not capable of growing at this temperature. One hundred percent of the C. dubliniensis isolates were unable to assimilate XYL and MDG when tested with the API 20C AUX system. However, according to the Vitek system, two (3.0%) and three (4.5%) separate isolates of C. dubliniensis were capable of assimilating XYL and MDG, respectively.

Among the 100 C. albicans isolates tested, 96 were able to assimilate XYL contained in both the API 20C AUX and Vitek systems. Four C. albicans strains (20052012, 20052061, 20052087, and 2-7IA) were not capable of assimilating the XYL as the sole carbon source in both systems (Table 1). The XYL tests were repeated, confirming the negative results. Three of the four XYL-negative C. albicans isolates assimilated MDG in the API 20C AUX system, but all assimilated MDG in the Vitek system. Three of the four XYL-negative C. albicans isolates did not grow at 45°C. The four C. albicans strains with negative XYL tests were reidentified by hybridization tests with the C. albicans-specific probe Ca3, verifying that they were C. albicans. The assimilation of MDG by the C. albicans isolates was 100 and 54% when they were tested with the Vitek and API 20C AUX systems, respectively.

With the C. dubliniensis-specific probe Cd25 used as the "gold standard" for identification, the sensitivities and specificities of the XYL and MDG tests performed by the API 20C AUX and Vitek systems for identification of the C. dubliniensis isolates were calculated (Table 2). A positive test for C. dubliniensis indicated that there was no assimilation of XYL and MDG as sources of carbon. Both systems showed a sensitivity of greater than 95.0% for identification of the C. dubliniensis isolates. However, the API 20C AUX system demonstrated a sensitivity (100.0%) superior to that of Vitek system for both tests. On the other hand, although both systems displayed the same specificity for XYL (96.0%), the MDG contained in the Vitek system was the only test with 100.0% specificity for detection of the C. dubliniensis isolates. Although the MDG sensitivity in the API 20C AUX system was excellent, its specificity was only 54.0%.

In order to evaluate the reproducibility of the XYL and MDG tests contained in the API trays and Vitek cards, 10 C. dubliniensis isolates and 10 C. albicans isolates were selected randomly for retesting. The results of the reproducibility tests are shown in Table 3. The API 20C AUX showed excellent reproducibility of the XYL and MDG tests for C. dubliniensis and of the XYL test for C. albicans. The combined qualitative accuracy of the tests was 100.0%. However, three variations were noticed for MDG when C. albicans isolates were tested in the API 20C AUX system. The reproducibility of the Vitek system was slightly lower than that of the API 20C AUX system for C. dubliniensis. Two and four variations of XYL and MDG results, respectively, were detected among the C. dubliniensis isolates with the Vitek system. Among the C. albicans strains, only one variation was detected in each test.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The ability to differentiate readily between isolates of C. dubliniensis and C. albicans in the routine diagnostic laboratory remains a technical problem. The CHROMagar Candida medium has been useful only for the presumptive identification of C. dubliniensis on primary isolation. On this medium colonies of C. dubliniensis demonstrate a dark green color on primary isolation, while C. albicans colonies show a light blue-green. However, it has been reported that this morphologic characteristic may be lost on repeated subculture or storage (7, 19). For this reason, we did not attempt to differentiate C. albicans from C. dubliniensis based on colony color on CHROMagar Candida, because the organisms studied were stored in our bank collection. However, Koehler and colleagues recently reported that one C. dubliniensis isolate retained its ability to form dark green colonies after repeated subculture (8).

Since Pinjon and colleagues reported that the ability to grow at 45°C could constitute a simple, inexpensive, and reliable method for differentiation of C. dubliniensis from C. albicans (17), this test has been used as a screening test in detection of C. dubliniensis isolates (5). Our results confirmed that this simple test could be used as a screening test in the detection of C. dubliniensis. Although none of the C. dubliniensis isolates was able to grow at 45°C, 23 of the C. albicans isolates tested had poor or no growth at this temperature as well. This demonstrates that this test has a low specificity and that C. albicans isolates could be falsely identified as C. dubliniensis. Kirkpatrick and colleagues observed similar results (7). Perhaps the specificity of the test could be increased if the isolates were subcultured in potato dextrose broth instead of potato dextrose agar and their growth was measured with a spectrophotometer. We did not analyze our isolates spectrophotometrically because we observed that growth at 45°C was not a reproducible test for C. albicans isolates; i.e., isolates that showed no growth the first time when subcultured demonstrated excellent growth in subsequent tests.

When Sullivan and colleagues described C. dubliniensis as a new species, they reported that this species was not able to assimilate XYL and MDG (22). Other studies have also reported that isolates of C. dubliniensis are not able to assimilate XYL and MDG with the API 20C AUX system (7, 18, 19). A similar evaluation of the Vitek system has not been performed. The differentiation of C. dubliniensis from C. albicans isolates based on assimilation of sugars in the API 20C AUX system has been difficult because occasional strains of C. albicans fail to assimilate XYL and MDG as well. In spite of these observations, we challenged the performance of the XYL and MDG tests as determined with the Vitek and API 20C AUX systems in order to distinguish C. dubliniensis isolates from C. albicans isolates. In our study, we also observed that isolates of C. dubliniensis were not able to assimilate XYL and MDG in the API 20C AUX system. According to the Vitek system, only two C. dubliniensis strains showed XYL-positive tests, while three C. dubliniensis isolates assimilated MDG (false-positive reactions). The XYL test contained in the API 20C AUX system was able to distinguish all 66 C. dubliniensis isolates from the 100 C. albicans isolates tested in this study. Despite a sensitivity of 100.0% for detection of C. dubliniensis isolates with the MDG test contained in the API 20C AUX system, this test proved to be useless for distinction between these two species because of the high number of C. albicans isolates that failed to assimilate MDG.

The XYL and MDG tests contained in the Vitek system showed a performance slightly inferior to that of the API 20C AUX tests for detection of C. dubliniensis isolates. However, among the C. albicans isolates, the XYL test showed the same performance exhibited by the API 20C AUX system. In addition, the Vitek MDG test demonstrated 100.0% specificity for detection of C. dubliniensis isolates, indicating that MDG contained in the Vitek system was an excellent test to differentiate the C. dubliniensis isolates from the C. albicans isolates.

Among the four XYL-negative C. albicans strains, only one isolate was differentiated from C. dubliniensis by growth at 45°C. When the MDG tests contained in the API 20C AUX and Vitek systems were evaluated, three of XYL-negative C. albicans isolates were classified correctly by the API 20C AUX system, and all four were classified correctly by the Vitek system.

The reproducibility of the tests was slightly better with the API 20C AUX system than with the Vitek system. Previous studies have reported difficulties in the reading of the API 20C AUX trays (1, 5, 19). It has been reported that variations in the density of the inoculum may produce false-positive or false-negative tests. False-positive assimilation reactions could be due to a heavy inoculum and would result in a carryover phenomenon where growth is observed in the zero-growth control cupule (2). In order to avoid false-positive or false-negative results, the inoculum must be adjusted to the proper density. Moreover, reading of the API 20C AUX assimilation results requires experience.

In comparison with the API 20C AUX system, the Vitek system is easier to use and less time-consuming. The Vitek system also offers early results (maximum of 48 h) and objective reading. Laboratories that already utilize the Vitek or API 20C AUX system as an identification method may suspect the presence of C. dubliniensis if the isolate being tested fails to assimilate XYL and MDG. However, the use of these systems for testing all clinical isolates of Candida is not practical in most clinical laboratories.

We conclude that the XYL test contained in both systems and the Vitek MDG test constitute useful tests for the distinction between C. dubliniensis and C. albicans. Negative XYL and MDG tests determined with these commercially available systems strongly suggest that an isolate is C. dubliniensis rather than C. albicans. The distinction between C. dubliniensis and C. albicans remains a challenge for clinical microbiology laboratories. No single phenotypic test has proven to be highly effective, and genotypic tests may be necessary for definitive identification.