CHROMagar Candida Medium for Direct Susceptibility Testing of Yeast from Blood Cultures
http://www.100md.com
微生物临床杂志 2005年第4期
Department of Pathology, Medical Microbiology Division, University of California Irvine, Medical Center, Orange, California 92868
ABSTRACT
An evaluation was performed on 95 blood cultures positive for Candida spp. to determine the correlation of direct susceptibility testing of fluconazole versus both standardized disk diffusion and MIC methods. For direct testing, an aliquot taken from BD BACTEC Plus and/or BD BACTEC Lytic/10 bottles (Becton Dickinson [BD], Sparks, MD) positive by gram stain for yeast was subcultured to CHROMagar Candida (BD), and a 25-μg fluconazole disk (BD) was placed on the plate. The area of growth inhibition surrounding the disk was measured at 24 and 48 h. In addition, a subculture of the isolate was tested by a microdilution MIC using YeastOne (TREK Diagnostics Systems Inc., OH) and disk diffusion (NCCLS M44-A) using a standardized inoculum plated onto CHROMagar Candida as well as Mueller-Hinton agar to which 2% glucose and 0.5 μg/ml methylene blue dye was added (MH-GMB). The categorical interpretation derived from the MIC was used as the reference to which the disk diffusion results were compared. There were a total of 41 Candida albicans, 23 Candida glabrata, 20 Candida parapsilosis, 9 Candida tropicalis, and 1 each of Candida krusei and Candida lusitaniae tested. At 24 h there was full agreement among the methods for all C. albicans, C. tropicalis, C. lusitaniae, and C. krusei isolates. For the C. parapsilosis isolates at 24 h there was one very major discrepancy using the direct CHROMagar and one major error with the standardized MH-GMB. The majority of the errors were seen at 24 h with the C. glabrata isolates. Of the 23 C. glabrata isolates at 24 h by direct CHROMagar, there were 10 minor and 1 very major error; by MH-GMB there were 12 minor and 2 very major errors; and by standardized CHROMagar Candida there were 13 minor and 2 major errors. There were no very major errors with C. glabrata when all plates were read at 48 h. At 24 h by the direct and standardized CHROMagar the majority of C. glabrata isolates were more resistant, whereas by MH-GMB they were more susceptible than the reference MIC interpretation. In summary, subculturing yeast directly from blood cultures onto CHROMagar to which a fluconazole disk has been added may provide a presumptive identification at 24 h and, with the exception of C. glabrata, was able to predict the susceptibility to fluconazole with the majority of Candida isolates examined in this evaluation.
INTRODUCTION
Over the past decade there has been more interest and subsequently progress in developing reliable methods for the susceptibility testing of yeast. In particular, testing Candida species against azole antifungal agents has provided valuable information for treatment of patients with invasive yeast infections. However, even with the advent of standardized guidelines for microdilution methods, testing is not readily available for a majority of laboratories due to the expense and labor that are inherent in the method (5, 6). Recently, guidelines for testing by disk diffusion have been published (7). This method utilizes a medium and a procedure that are familiar to the vast majority of laboratories, thereby making it more practical for implementation.
Patients with a systemic yeast infection often have a positive blood culture. In these cases rapid feedback to the physician as to the identification of the yeast provides valuable information concerning treatment regimens (5). For instance, while Candida albicans is susceptible to fluconazole, Candida glabrata isolates are frequently resistant to this antifungal agent (1). CHROMagar Candida medium (CHROMagar) has been reported to achieve the goal of rapid and reliable direct isolation and in some cases identification of Candida species (4). To date, however, there have been no published studies to evaluate the use of CHROMagar for direct susceptibility testing. In this study we examined the ability of CHROMagar to identify Candida to the species level and to predict the susceptibility to fluconazole by direct plating and disk diffusion susceptibility testing from blood cultures positive for yeast.
MATERIALS AND METHODS
Isolates. Isolates used in this evaluation were obtained from blood cultures submitted to the Medical Microbiology Laboratory at the University of California Irvine Medical Center. Blood was collected in BD BACTEC Plus and/or BD BACTEC Lytic/10 bottles (Becton Dickinson [BD], Sparks, MD). The work performed in this study was approved by the Institutional Review Board at the University of California, Irvine.
In addition to fresh clinical isolates, yeast isolates frozen in fetal bovine serum with 10% glycerol were examined. Yeasts were thawed and subcultured on Sabouraud dextrose agar twice before being tested. A standardized suspension of yeast was made from an overnight culture on Sabouraud dextrose agar in order to inoculate 1,000 CFU into blood cultures that were negative after 5 days. Seeded cultures were then monitored by the BACTEC 9240 system, and upon being flagged as positive an aliquot was used as described below for direct and standardized testing of susceptibility to fluconazole.
Direct susceptibility. Upon being flagged positive by the BACTEC 9240 (BD), cultures containing yeast by gram stain were inoculated to CHROMagar Candida (BD). A needle and syringe were used to remove an aliquot from the blood culture bottle, and 0.2 ml was placed on the agar surface. Subsequently, a cotton swab was used to spread the inoculum over one third of the plate in order to have a confluent area of growth. A loop was then used to streak for isolation using the remainder of the agar surface. A 25-μg fluconazole disk (BD) was then placed on the swabbed area of the plate, and the plate was incubated for 24 h at 35°C in air. The zone of inhibition around the disk was read to the point at which there was a prominent reduction in growth. Plates were reincubated and read again at 48 h.
MIC. The reference method used for this study was the categorical interpretation of the MIC obtained using YeastOne (TREK Diagnostic Systems, Cleveland, Ohio). Plates were inoculated and read using manufacturer's instructions. Briefly, isolates from a 24-h culture on Sabouraud dextrose agar (BD) were suspended in water to the equivalent of a 0.5 McFarland turbidity standard. The final inoculum was adjusted to 1.5 x 103 to 8 x 103 CFU/ml. Plates were incubated at 35°C in air, and the MIC was read as the lowest concentration of fluconazole that substantially inhibited the growth of the yeast as detected by observing a color change. Interpretation of the MIC was determined by the values given in NCCLS document M27-A2 (6). Quality control was performed as stated by the manufacturer using C. parapsilosis (ATCC 22019).
Standardized disk diffusion. For a standardized method for disk diffusion, 100-mm Mueller-Hinton agar plates (BD) to which 1.5 ml of 2% glucose and 0.5 μg/ml methylene blue dye was added (MH-GMB) to the surface of the agar as described for NCCLS M44-A and CHROMagar (BD) were used. As described above, a standardized 0.5 McFarland suspension of yeast taken from Sabouraud dextrose agar was made in 0.85% saline. Using a sterile swab, plates were inoculated to produce a confluent lawn of growth and a fluconazole disk (25 μg) applied. Plates were inverted and incubated for 24 h at 35°C. Plates were read at 24 and 48 h. Zones of inhibition were read as described above for the direct testing, and interpretation of zone diameters was as described in NCCLS M44-A (7). Quality control was performed using C. parapsilosis (ATCC 22019).
Identification. Isolates were subcultured to Sabouraud dextrose agar and identified by standard methods, including germ tube production, microscopic morphology on cornmeal agar with Polysorbate 80 (BD), and a yeast biochemical identification card (bioMerieux sa, Durham, North Carolina) (3). Isolates giving an identification of <90% by the identification card were further tested using API 20 C (bioMerieux sa).
Data analysis. The interpretive categories used were susceptible (S), susceptible-dose dependent (S-DD), and resistant (R). For purposes of data analysis the following definitions were used: minor errors occurred when one of the methods was S-DD and the other was either S or R; a major error was defined when the reference method was S and the disk diffusion R; a very major error was one in which the reference was R and the disk diffusion method was S.
RESULTS
Isolate distribution. A total of 95 blood cultures positive for Candida species from 95 patients were included in the evaluation. The distribution of isolates was as follows: C. albicans, 41; C. glabrata, 23; C. parapsilosis, 20; C. tropicalis, 9; C. krusei, 1; and C. lusitaniae, 1. Using the categorical interpretation derived from the MIC, all C. albicans, C. tropicalis, and C. lusitaniae isolates were susceptible (Table 1). Of the 23 C. glabrata isolates examined, 3 were susceptible, 14 susceptible-dose dependent, and 6 resistant. All C. parapsilosis isolates were susceptible except for two isolates, one of which was susceptible-dose dependent and one of which was resistant. The one C. krusei isolate was resistant, as expected due to C. krusei's intrinsic resistance to fluconazole (5).
Direct CHROMagar. When the zone of inhibition surrounding the fluconazole was measured at 24 h from the CHROMagar directly inoculated from the positive blood culture, all C. albicans, C. tropicalis, and C. lusitaniae isolates were susceptible, agreeing with the MIC interpretation (Tables 1 and 2). In addition, the one C. krusei isolate was resistant, also agreeing with the reference interpretation. The interpretation of the zone diameters for these isolates did not change when read at 48 h, with the exception of one C. tropicalis isolate, which gave a smaller zone size, making it susceptible-dose dependent.
Of the 20 C. parapsilosis isolates, 18 agreed with the MIC interpretation and one isolate yielded a minor error, being susceptible-dose dependent by the direct CHROMagar and susceptible by the reference. There was a very major error with the direct CHROMagar with one C. parapsilosis isolate that was resistant by the reference method yet susceptible by the 24-h direct CHROMagar reading (Table 2). When the zone of inhibition was read at 48 h, this isolate was clearly resistant by the direct CHROMagar.
Of the 23 C. glabrata isolates, 12 agreed (2 susceptible, 5 susceptible-dose dependent and 5 resistant) with the reference method when the direct CHROMagar was read at 24 h (Table 2). Ten of the C. glabrata isolates which were susceptible-dose dependent by the reference method yielded minor errors by the direct CHROMagar when read at 24 h. Seven of these isolates were resistant and three were susceptible by the direct method. There was a very major error with one of the C. glabrata isolates at 24 h. Reading the direct CHROMagar at 48 h eliminated the one very major error; however, the minor errors were increased from 10 to 12. In addition, reading plates at 48 h introduced a major error, where the zone of inhibition of an isolate susceptible by the reference decreased enough, from 18 mm to 14 mm by 48 h, that it was interpreted as resistant.
Overall, reading the direct CHROMagar at 24 h yielded complete agreement with the reference interpretation for 86% of the isolates (82/95) and very major discrepancies for 2% (2/95) of the isolates, whereas reading of the zones of inhibition at 48 h decreased the overall agreement to 83% (79/95) but eliminated the two very major errors found at 24 h (Table 3).
Standardized CHROMagar. Similar findings at 24 h as the direct CHROMagar result were found when a standardized suspension of Candida of the blood culture isolates recovered in this evaluation was used to inoculate CHROMagar to test for fluconazole susceptibility. When the zone diameters were read at 24 h, all C. albicans, C. tropicalis, C. lusitaniae, and C. krusei isolates exhibited the same result as the reference method (Table 2). With the C. parapsilosis isolates, at 24 h all readings except one, which gave a minor error, agreed with the reference result. Readings taken at 48 h yielded the same interpretation except for one C. tropicalis and one C. parapsilosis isolate, which went from susceptible at 24 h to susceptible-dose dependent.
With the 23 C. glabrata isolates, there were no very major errors and 8 agreed with the reference interpretation at 24 h. There were 13 minor errors at 24 h, largely due to many of the isolates which were susceptible dose-dependent by the reference method yet resistant on CHROMagar. There were two major errors at 24 h that remained at the 48-h reading.
Overall, the standardized CHROMagar gave 83% (79/95) and 80% (76/95) agreement, respectively, for the readings taken at 24 h and 48 h compared to the MIC interpretation (Table 3). There were no very major errors, while 2% (2/95) of the isolates tested exhibited a major error. There were 15% (14/95) and 18% (17/95) of the isolates with minor errors. In general, the minor errors were due to a more resistant reading from the standardized CHROMagar than the YeastOne reference method.
Standardized MH-GMB. There was complete agreement with the reference interpretation when the C. albicans, C. tropicalis, C. lusitaniae, and C. krusei were tested using the MH-GMB agar as outlined in NCCLS M44-A (7). Reading the zone of inhibition at 24 h and 48 h around the fluconazole disk yielded the same interpretation (Tables 1 and 2).
As with both the direct and standardized CHROMagar, discrepancies were seen with the reference interpretation with isolates of C. parapsilosis and C. glabrata. At the 24-h reading with C. parapsilosis isolates, all agreed except two isolates, one which exhibited a minor and one a major discrepancy. At 48 h an additional major error was detected (Table 2).
Of the C. glabrata isolates when zones of inhibition were read at 24 h there were two very major discrepancies seen. These resolved when the zones were read at 48 h, at which time only minor errors were detected. This is in part due to the difficulty encountered in reading the C. glabrata zones at 24 h, most likely due to their slower growth. Interestingly, as with the direct and standardized CHROMagar readings at 24 h there were several (n = 12) minor errors. However, with MH-GMB these minor errors were due to a more susceptible reading in contrast to CHROMagar, where the readings resulting in a minor error were more resistant than the reference method (Table 1). When these initially susceptible isolates by MH-GMB were read at 48 h, they were interpreted as resistant, agreeing with the initial CHROMagar reading. Assuming a similar growth rate on the two media and knowing the same standardized inoculum was used for both, it may be that the chromogenic substrates in the CHROMagar allowed reduced growth within the zone to be detected earlier, allowing the reduced susceptibility to be detected by 24 h.
Overall, MH-GMB agar for the 24 h result gave the same percentage agreement, 83% (79/95), as the standardized CHROMagar, while the 48-h reading, 86% (82/95), was slightly higher than that achieved with the standardized CHROMagar. As with the direct CHROMagar, 2% (2/95) of the isolates at 24 h gave a very major error, and both errors were resolved with the 48-h reading. At 24 and 48 h, respectively, 1% and 2% of the isolates showed a major discrepancy. The 14% (13/95) minor discrepancies at 24 h were due to a more susceptible interpretation than that of the reference MIC. This was the exact opposite of both the direct and standardized CHROMagar result, where the minor errors were due to smaller zones resulting in a more resistant interpretation.
Seeded isolates. Since it was clear that the majority of discrepancies were seen with the more resistant isolates, namely C. glabrata, to expand our data and also to test for reproducibility of the direct CHROMagar we seeded known negative blood cultures with 10 clinical isolates of C. glabrata that had been frozen at –70°C. Of these, three were susceptible by the reference method, five were susceptible-dose dependent, and two were read as resistant. This subset produced similar results to those reported with the fresh clinical isolates. Here six isolates agreed with the reference when read at 24 h from both the direct and standardized CHROMagar, with three minor and one major discrepancy. The standardized MH-GMB yielded five isolates that showed agreement with the reference and five with minor errors. As with the fresh isolates, the minor errors at 24 h were due to more resistant readings with the CHROMagar, both direct and standardized, and more susceptible readings from the MH-GMB.
Once seeded blood culture bottles turned positive they were subcultured in duplicate to test the reproducibility of the zone diameters. Here, at the 24-h reading, 7 of the 10 isolates yielded essentially the same zone size, i.e., within 1 mm of each other, and at 48 h 8 were read the same. At 24 h there were two isolates with duplicate plates yielding zones of inhibition within 2 mm of one another and one with a 4-mm difference. Even though there were some differences in the zone diameters recorded, none of the interpretations changed.
Identification from direct CHROMagar. Using characteristic colony color and/or size and the size of the individual yeast, identification of all isolates of C. albicans, C. glabrata, C. tropicalis, and C. krusei from the 24-h direct CHROMagar agreed 100% with those obtained with conventional identification methods. In some cases, in order to detect their characteristic colony color on CHROMagar, isolates of C. glabrata needed to be incubated for 48 h. Due to a lack of distinguishing features on CHROMagar the C. lusitaniae and 20 C. parapsilosis isolates were unable to be identified from CHROMagar.
DISCUSSION
In this study we present evidence that positive blood cultures containing yeast subcultured directly onto CHROMagar allows the rapid identification and determination of fluconazole susceptibility for the majority of Candida isolates encountered in the clinical laboratory. Of the 95 Candida isolates reported in this series, 78% (74/95) were identified from the direct CHROMagar subcultures. In addition, when a fluconazole disk was placed directly on the CHROMagar, 86% (82/95) of isolates gave the same categorical interpretation as that obtained by the reference MIC. Therefore, using CHROMagar to directly subculture blood culture bottles positive for yeast has the potential to provide rapid clinically relevant information.
Our findings on the direct identification of Candida from blood cultures using CHROMagar corroborate those previously reported by Horvath et al. (4). Using 50 seeded blood cultures they were able to correctly identify all isolates of C. albicans (n = 12), C. tropicalis (n = 12), C. glabrata (n = 9), and C. krusei (n = 5). Our study differed from theirs in that we used the actual patient blood culture, whereas they seeded stock isolates into blood cultures.
It is clear from our data there is a tendency to undercall the level of resistance with C. glabrata when the standardized MH-GMB plate is read at 24 h. Therefore, we recommend routinely extending the incubation of the fluconazole disk diffusion test to 48 h for all C. glabrata isolates. In doing so the agreement with the reference method nearly doubled from 39% to 61%. Pfaller et al. (8) in evaluating fluconazole susceptibility testing with 235 C. glabrata isolates reported a 64.7% categorical agreement when comparing MH-GMB agar disk diffusion results with a reference MIC interpretation. They reported, as we have seen, that the MH-GMB disk diffusion method was unable to distinguish between S and S-DD at 24 h, resulting in minor errors. They did not report results of a 48-h reading. The results of their 24-h reading were actually similar to ours at 48 h. The reason for this is unclear. It could be due to the differences in the media and reference methods used in our respective studies. We used YeastOne as the MIC reference method, while they prepared microtiter MIC plates in their laboratory. Espinel-Ingroff et al. (2) compared the performance of the YeastOne panel used here to the NCCLS M27-A broth microdilution method (6). They found a 98 to 95% agreement with the non-C. albicans isolates, while the C. albicans isolates had an 87% agreement between methods. We added GMB to premade MH plates, whereas in the study conducted by Pfaller et al. (8) it was not indicated whether they used media with GMB incorporated into the initial agar pour or added the GMB to the surface of prepared MH plates. At this point, we do not know the reason for the differences in the zone sizes observed with C. glabrata when measured on MH-GMB versus CHROMagar, with smaller zone sizes being seen on CHROMagar. This could be due to slower growth of C. glabrata on MH-GMB, resulting in a larger zone size and/or color production on CHROMagar, which may actually aid in visualizing the smaller colonies inside the zone of inhibition, resulting in a smaller zone size.
Reading the zone of inhibition around the fluconazole disk is at times difficult and quite subjective. With yeast, the zone of inhibition is read where there is a predominant reduction of growth, unlike the conventional Kirby-Bauer method used for most bacterium-antimicrobial combinations, where the zone of inhibition is distinct (7). Nevertheless, overall, interpretation of results obtained from disk diffusion correlated >80% of the time with the categorical interpretation derived from the reference method. The two very major errors seen with the more resistant isolates of C. glabrata and C. parapsilosis are troubling. However, by extending the incubation time to 48 h, the very major errors seen with these two species were resolved. When using the direct method, extension of the incubation time to 48 h for these two species is recommended. This is also true of the NCCLS M44-A method using MH-GMB, which revealed the same problem when fluconazole zones were read at 24 h with C. glabrata isolates. In fact, the problem was even greater using the method recommended in NCCLS M44-A (7) in comparison to the direct method using CHROMagar described in this report. In summary, use of CHROMagar allowed rapid identification of commonly isolated yeasts from blood cultures and also provided information on the susceptibility to fluconazole to aid in treatment decisions.
REFERENCES
Antoniadou, A., H. A. Torres, R. E. Lewis, J. Thornby, G. P. Bodey, J. P. Tarrand, X. Y. Han, K. V. Rolston, A. Safdar, I. I. Raad, and D. P. Kontoyiannis. 2003. Candidemia in a tertiary care cancer center: in vitro susceptibility and its association with outcome of initial antifungal therapy. Medicine 82:309-321.
Espinel-Ingroff, A., M. Pfaller, S. A. Medsser, C. C. Knapp, S. Killian, H. A. Norris, and M. A. Ghannoum. 1999. Multicenter comparison of the Sensititre YeastOne colorimetric antifungal panel with the National Committee for Clinical Laboratory Standards M27-A reference method for testing clinical isolates of common and emerging Candida spp., Cryptococcus spp., and other yeasts and yeast-like organisms. J. Clin. Microbiol. 37:591-595.
Hazen, K., and S. A. Howell. 2003. Candida, Cryptococcus, and other yeasts of medical importance, p. 1693-1711. In P. M. Murray, E. J. Baron, J. H. Jorgensen, M. A. Pfaller, and R. H. Yolken (ed.), Manual of clinical microbiology, 8th ed. ASM Press, Washington, D.C.
Horvath, L. L., D. R. Hospenthal, C. K. Murray, and D. P. Dooley. 2003. Direct isolation of Candida spp. from blood cultures on the chromogenic medium CHROMagar Candida. J. Clin. Microbiol. 41:2629-2632.
Hospenthal, D. R., C. K. Murray, and M. G. Rinaldi. 2004. The role of antifungal susceptibility testing in the therapy of candidiasis. Diagn. Microbiol. Infect. Dis. 48:153-160.
National Committee for Clinical Laboratory Standards. 2002. Reference method for broth dilution antifungal susceptibility testing of yeasts. Approved standard M27-A2. National Committee for Clinical Laboratory Standards, Wayne, Pa.
National Committee for Clinical Laboratory Standards. 2004. Method for antifungal disk diffusion susceptibility testing of yeasts: proposed guideline M44-A. National Committee for Clinical Laboratory Standards, Wayne, Pa.
Pfaller, M. A., D. J. Diekema, L. Boyken, S. A. Messer, S. Tendolkar, and R. J. Hollis. 2003. Evaluation of the Etest and disk diffusion methods for determining susceptibilities of 235 bloodstream isolates of Candida glabrata to fluconazole and voriconazole. J. Clin. Microbiol. 41:1875-1880.(Grace L. Tan and Ellena M)
ABSTRACT
An evaluation was performed on 95 blood cultures positive for Candida spp. to determine the correlation of direct susceptibility testing of fluconazole versus both standardized disk diffusion and MIC methods. For direct testing, an aliquot taken from BD BACTEC Plus and/or BD BACTEC Lytic/10 bottles (Becton Dickinson [BD], Sparks, MD) positive by gram stain for yeast was subcultured to CHROMagar Candida (BD), and a 25-μg fluconazole disk (BD) was placed on the plate. The area of growth inhibition surrounding the disk was measured at 24 and 48 h. In addition, a subculture of the isolate was tested by a microdilution MIC using YeastOne (TREK Diagnostics Systems Inc., OH) and disk diffusion (NCCLS M44-A) using a standardized inoculum plated onto CHROMagar Candida as well as Mueller-Hinton agar to which 2% glucose and 0.5 μg/ml methylene blue dye was added (MH-GMB). The categorical interpretation derived from the MIC was used as the reference to which the disk diffusion results were compared. There were a total of 41 Candida albicans, 23 Candida glabrata, 20 Candida parapsilosis, 9 Candida tropicalis, and 1 each of Candida krusei and Candida lusitaniae tested. At 24 h there was full agreement among the methods for all C. albicans, C. tropicalis, C. lusitaniae, and C. krusei isolates. For the C. parapsilosis isolates at 24 h there was one very major discrepancy using the direct CHROMagar and one major error with the standardized MH-GMB. The majority of the errors were seen at 24 h with the C. glabrata isolates. Of the 23 C. glabrata isolates at 24 h by direct CHROMagar, there were 10 minor and 1 very major error; by MH-GMB there were 12 minor and 2 very major errors; and by standardized CHROMagar Candida there were 13 minor and 2 major errors. There were no very major errors with C. glabrata when all plates were read at 48 h. At 24 h by the direct and standardized CHROMagar the majority of C. glabrata isolates were more resistant, whereas by MH-GMB they were more susceptible than the reference MIC interpretation. In summary, subculturing yeast directly from blood cultures onto CHROMagar to which a fluconazole disk has been added may provide a presumptive identification at 24 h and, with the exception of C. glabrata, was able to predict the susceptibility to fluconazole with the majority of Candida isolates examined in this evaluation.
INTRODUCTION
Over the past decade there has been more interest and subsequently progress in developing reliable methods for the susceptibility testing of yeast. In particular, testing Candida species against azole antifungal agents has provided valuable information for treatment of patients with invasive yeast infections. However, even with the advent of standardized guidelines for microdilution methods, testing is not readily available for a majority of laboratories due to the expense and labor that are inherent in the method (5, 6). Recently, guidelines for testing by disk diffusion have been published (7). This method utilizes a medium and a procedure that are familiar to the vast majority of laboratories, thereby making it more practical for implementation.
Patients with a systemic yeast infection often have a positive blood culture. In these cases rapid feedback to the physician as to the identification of the yeast provides valuable information concerning treatment regimens (5). For instance, while Candida albicans is susceptible to fluconazole, Candida glabrata isolates are frequently resistant to this antifungal agent (1). CHROMagar Candida medium (CHROMagar) has been reported to achieve the goal of rapid and reliable direct isolation and in some cases identification of Candida species (4). To date, however, there have been no published studies to evaluate the use of CHROMagar for direct susceptibility testing. In this study we examined the ability of CHROMagar to identify Candida to the species level and to predict the susceptibility to fluconazole by direct plating and disk diffusion susceptibility testing from blood cultures positive for yeast.
MATERIALS AND METHODS
Isolates. Isolates used in this evaluation were obtained from blood cultures submitted to the Medical Microbiology Laboratory at the University of California Irvine Medical Center. Blood was collected in BD BACTEC Plus and/or BD BACTEC Lytic/10 bottles (Becton Dickinson [BD], Sparks, MD). The work performed in this study was approved by the Institutional Review Board at the University of California, Irvine.
In addition to fresh clinical isolates, yeast isolates frozen in fetal bovine serum with 10% glycerol were examined. Yeasts were thawed and subcultured on Sabouraud dextrose agar twice before being tested. A standardized suspension of yeast was made from an overnight culture on Sabouraud dextrose agar in order to inoculate 1,000 CFU into blood cultures that were negative after 5 days. Seeded cultures were then monitored by the BACTEC 9240 system, and upon being flagged as positive an aliquot was used as described below for direct and standardized testing of susceptibility to fluconazole.
Direct susceptibility. Upon being flagged positive by the BACTEC 9240 (BD), cultures containing yeast by gram stain were inoculated to CHROMagar Candida (BD). A needle and syringe were used to remove an aliquot from the blood culture bottle, and 0.2 ml was placed on the agar surface. Subsequently, a cotton swab was used to spread the inoculum over one third of the plate in order to have a confluent area of growth. A loop was then used to streak for isolation using the remainder of the agar surface. A 25-μg fluconazole disk (BD) was then placed on the swabbed area of the plate, and the plate was incubated for 24 h at 35°C in air. The zone of inhibition around the disk was read to the point at which there was a prominent reduction in growth. Plates were reincubated and read again at 48 h.
MIC. The reference method used for this study was the categorical interpretation of the MIC obtained using YeastOne (TREK Diagnostic Systems, Cleveland, Ohio). Plates were inoculated and read using manufacturer's instructions. Briefly, isolates from a 24-h culture on Sabouraud dextrose agar (BD) were suspended in water to the equivalent of a 0.5 McFarland turbidity standard. The final inoculum was adjusted to 1.5 x 103 to 8 x 103 CFU/ml. Plates were incubated at 35°C in air, and the MIC was read as the lowest concentration of fluconazole that substantially inhibited the growth of the yeast as detected by observing a color change. Interpretation of the MIC was determined by the values given in NCCLS document M27-A2 (6). Quality control was performed as stated by the manufacturer using C. parapsilosis (ATCC 22019).
Standardized disk diffusion. For a standardized method for disk diffusion, 100-mm Mueller-Hinton agar plates (BD) to which 1.5 ml of 2% glucose and 0.5 μg/ml methylene blue dye was added (MH-GMB) to the surface of the agar as described for NCCLS M44-A and CHROMagar (BD) were used. As described above, a standardized 0.5 McFarland suspension of yeast taken from Sabouraud dextrose agar was made in 0.85% saline. Using a sterile swab, plates were inoculated to produce a confluent lawn of growth and a fluconazole disk (25 μg) applied. Plates were inverted and incubated for 24 h at 35°C. Plates were read at 24 and 48 h. Zones of inhibition were read as described above for the direct testing, and interpretation of zone diameters was as described in NCCLS M44-A (7). Quality control was performed using C. parapsilosis (ATCC 22019).
Identification. Isolates were subcultured to Sabouraud dextrose agar and identified by standard methods, including germ tube production, microscopic morphology on cornmeal agar with Polysorbate 80 (BD), and a yeast biochemical identification card (bioMerieux sa, Durham, North Carolina) (3). Isolates giving an identification of <90% by the identification card were further tested using API 20 C (bioMerieux sa).
Data analysis. The interpretive categories used were susceptible (S), susceptible-dose dependent (S-DD), and resistant (R). For purposes of data analysis the following definitions were used: minor errors occurred when one of the methods was S-DD and the other was either S or R; a major error was defined when the reference method was S and the disk diffusion R; a very major error was one in which the reference was R and the disk diffusion method was S.
RESULTS
Isolate distribution. A total of 95 blood cultures positive for Candida species from 95 patients were included in the evaluation. The distribution of isolates was as follows: C. albicans, 41; C. glabrata, 23; C. parapsilosis, 20; C. tropicalis, 9; C. krusei, 1; and C. lusitaniae, 1. Using the categorical interpretation derived from the MIC, all C. albicans, C. tropicalis, and C. lusitaniae isolates were susceptible (Table 1). Of the 23 C. glabrata isolates examined, 3 were susceptible, 14 susceptible-dose dependent, and 6 resistant. All C. parapsilosis isolates were susceptible except for two isolates, one of which was susceptible-dose dependent and one of which was resistant. The one C. krusei isolate was resistant, as expected due to C. krusei's intrinsic resistance to fluconazole (5).
Direct CHROMagar. When the zone of inhibition surrounding the fluconazole was measured at 24 h from the CHROMagar directly inoculated from the positive blood culture, all C. albicans, C. tropicalis, and C. lusitaniae isolates were susceptible, agreeing with the MIC interpretation (Tables 1 and 2). In addition, the one C. krusei isolate was resistant, also agreeing with the reference interpretation. The interpretation of the zone diameters for these isolates did not change when read at 48 h, with the exception of one C. tropicalis isolate, which gave a smaller zone size, making it susceptible-dose dependent.
Of the 20 C. parapsilosis isolates, 18 agreed with the MIC interpretation and one isolate yielded a minor error, being susceptible-dose dependent by the direct CHROMagar and susceptible by the reference. There was a very major error with the direct CHROMagar with one C. parapsilosis isolate that was resistant by the reference method yet susceptible by the 24-h direct CHROMagar reading (Table 2). When the zone of inhibition was read at 48 h, this isolate was clearly resistant by the direct CHROMagar.
Of the 23 C. glabrata isolates, 12 agreed (2 susceptible, 5 susceptible-dose dependent and 5 resistant) with the reference method when the direct CHROMagar was read at 24 h (Table 2). Ten of the C. glabrata isolates which were susceptible-dose dependent by the reference method yielded minor errors by the direct CHROMagar when read at 24 h. Seven of these isolates were resistant and three were susceptible by the direct method. There was a very major error with one of the C. glabrata isolates at 24 h. Reading the direct CHROMagar at 48 h eliminated the one very major error; however, the minor errors were increased from 10 to 12. In addition, reading plates at 48 h introduced a major error, where the zone of inhibition of an isolate susceptible by the reference decreased enough, from 18 mm to 14 mm by 48 h, that it was interpreted as resistant.
Overall, reading the direct CHROMagar at 24 h yielded complete agreement with the reference interpretation for 86% of the isolates (82/95) and very major discrepancies for 2% (2/95) of the isolates, whereas reading of the zones of inhibition at 48 h decreased the overall agreement to 83% (79/95) but eliminated the two very major errors found at 24 h (Table 3).
Standardized CHROMagar. Similar findings at 24 h as the direct CHROMagar result were found when a standardized suspension of Candida of the blood culture isolates recovered in this evaluation was used to inoculate CHROMagar to test for fluconazole susceptibility. When the zone diameters were read at 24 h, all C. albicans, C. tropicalis, C. lusitaniae, and C. krusei isolates exhibited the same result as the reference method (Table 2). With the C. parapsilosis isolates, at 24 h all readings except one, which gave a minor error, agreed with the reference result. Readings taken at 48 h yielded the same interpretation except for one C. tropicalis and one C. parapsilosis isolate, which went from susceptible at 24 h to susceptible-dose dependent.
With the 23 C. glabrata isolates, there were no very major errors and 8 agreed with the reference interpretation at 24 h. There were 13 minor errors at 24 h, largely due to many of the isolates which were susceptible dose-dependent by the reference method yet resistant on CHROMagar. There were two major errors at 24 h that remained at the 48-h reading.
Overall, the standardized CHROMagar gave 83% (79/95) and 80% (76/95) agreement, respectively, for the readings taken at 24 h and 48 h compared to the MIC interpretation (Table 3). There were no very major errors, while 2% (2/95) of the isolates tested exhibited a major error. There were 15% (14/95) and 18% (17/95) of the isolates with minor errors. In general, the minor errors were due to a more resistant reading from the standardized CHROMagar than the YeastOne reference method.
Standardized MH-GMB. There was complete agreement with the reference interpretation when the C. albicans, C. tropicalis, C. lusitaniae, and C. krusei were tested using the MH-GMB agar as outlined in NCCLS M44-A (7). Reading the zone of inhibition at 24 h and 48 h around the fluconazole disk yielded the same interpretation (Tables 1 and 2).
As with both the direct and standardized CHROMagar, discrepancies were seen with the reference interpretation with isolates of C. parapsilosis and C. glabrata. At the 24-h reading with C. parapsilosis isolates, all agreed except two isolates, one which exhibited a minor and one a major discrepancy. At 48 h an additional major error was detected (Table 2).
Of the C. glabrata isolates when zones of inhibition were read at 24 h there were two very major discrepancies seen. These resolved when the zones were read at 48 h, at which time only minor errors were detected. This is in part due to the difficulty encountered in reading the C. glabrata zones at 24 h, most likely due to their slower growth. Interestingly, as with the direct and standardized CHROMagar readings at 24 h there were several (n = 12) minor errors. However, with MH-GMB these minor errors were due to a more susceptible reading in contrast to CHROMagar, where the readings resulting in a minor error were more resistant than the reference method (Table 1). When these initially susceptible isolates by MH-GMB were read at 48 h, they were interpreted as resistant, agreeing with the initial CHROMagar reading. Assuming a similar growth rate on the two media and knowing the same standardized inoculum was used for both, it may be that the chromogenic substrates in the CHROMagar allowed reduced growth within the zone to be detected earlier, allowing the reduced susceptibility to be detected by 24 h.
Overall, MH-GMB agar for the 24 h result gave the same percentage agreement, 83% (79/95), as the standardized CHROMagar, while the 48-h reading, 86% (82/95), was slightly higher than that achieved with the standardized CHROMagar. As with the direct CHROMagar, 2% (2/95) of the isolates at 24 h gave a very major error, and both errors were resolved with the 48-h reading. At 24 and 48 h, respectively, 1% and 2% of the isolates showed a major discrepancy. The 14% (13/95) minor discrepancies at 24 h were due to a more susceptible interpretation than that of the reference MIC. This was the exact opposite of both the direct and standardized CHROMagar result, where the minor errors were due to smaller zones resulting in a more resistant interpretation.
Seeded isolates. Since it was clear that the majority of discrepancies were seen with the more resistant isolates, namely C. glabrata, to expand our data and also to test for reproducibility of the direct CHROMagar we seeded known negative blood cultures with 10 clinical isolates of C. glabrata that had been frozen at –70°C. Of these, three were susceptible by the reference method, five were susceptible-dose dependent, and two were read as resistant. This subset produced similar results to those reported with the fresh clinical isolates. Here six isolates agreed with the reference when read at 24 h from both the direct and standardized CHROMagar, with three minor and one major discrepancy. The standardized MH-GMB yielded five isolates that showed agreement with the reference and five with minor errors. As with the fresh isolates, the minor errors at 24 h were due to more resistant readings with the CHROMagar, both direct and standardized, and more susceptible readings from the MH-GMB.
Once seeded blood culture bottles turned positive they were subcultured in duplicate to test the reproducibility of the zone diameters. Here, at the 24-h reading, 7 of the 10 isolates yielded essentially the same zone size, i.e., within 1 mm of each other, and at 48 h 8 were read the same. At 24 h there were two isolates with duplicate plates yielding zones of inhibition within 2 mm of one another and one with a 4-mm difference. Even though there were some differences in the zone diameters recorded, none of the interpretations changed.
Identification from direct CHROMagar. Using characteristic colony color and/or size and the size of the individual yeast, identification of all isolates of C. albicans, C. glabrata, C. tropicalis, and C. krusei from the 24-h direct CHROMagar agreed 100% with those obtained with conventional identification methods. In some cases, in order to detect their characteristic colony color on CHROMagar, isolates of C. glabrata needed to be incubated for 48 h. Due to a lack of distinguishing features on CHROMagar the C. lusitaniae and 20 C. parapsilosis isolates were unable to be identified from CHROMagar.
DISCUSSION
In this study we present evidence that positive blood cultures containing yeast subcultured directly onto CHROMagar allows the rapid identification and determination of fluconazole susceptibility for the majority of Candida isolates encountered in the clinical laboratory. Of the 95 Candida isolates reported in this series, 78% (74/95) were identified from the direct CHROMagar subcultures. In addition, when a fluconazole disk was placed directly on the CHROMagar, 86% (82/95) of isolates gave the same categorical interpretation as that obtained by the reference MIC. Therefore, using CHROMagar to directly subculture blood culture bottles positive for yeast has the potential to provide rapid clinically relevant information.
Our findings on the direct identification of Candida from blood cultures using CHROMagar corroborate those previously reported by Horvath et al. (4). Using 50 seeded blood cultures they were able to correctly identify all isolates of C. albicans (n = 12), C. tropicalis (n = 12), C. glabrata (n = 9), and C. krusei (n = 5). Our study differed from theirs in that we used the actual patient blood culture, whereas they seeded stock isolates into blood cultures.
It is clear from our data there is a tendency to undercall the level of resistance with C. glabrata when the standardized MH-GMB plate is read at 24 h. Therefore, we recommend routinely extending the incubation of the fluconazole disk diffusion test to 48 h for all C. glabrata isolates. In doing so the agreement with the reference method nearly doubled from 39% to 61%. Pfaller et al. (8) in evaluating fluconazole susceptibility testing with 235 C. glabrata isolates reported a 64.7% categorical agreement when comparing MH-GMB agar disk diffusion results with a reference MIC interpretation. They reported, as we have seen, that the MH-GMB disk diffusion method was unable to distinguish between S and S-DD at 24 h, resulting in minor errors. They did not report results of a 48-h reading. The results of their 24-h reading were actually similar to ours at 48 h. The reason for this is unclear. It could be due to the differences in the media and reference methods used in our respective studies. We used YeastOne as the MIC reference method, while they prepared microtiter MIC plates in their laboratory. Espinel-Ingroff et al. (2) compared the performance of the YeastOne panel used here to the NCCLS M27-A broth microdilution method (6). They found a 98 to 95% agreement with the non-C. albicans isolates, while the C. albicans isolates had an 87% agreement between methods. We added GMB to premade MH plates, whereas in the study conducted by Pfaller et al. (8) it was not indicated whether they used media with GMB incorporated into the initial agar pour or added the GMB to the surface of prepared MH plates. At this point, we do not know the reason for the differences in the zone sizes observed with C. glabrata when measured on MH-GMB versus CHROMagar, with smaller zone sizes being seen on CHROMagar. This could be due to slower growth of C. glabrata on MH-GMB, resulting in a larger zone size and/or color production on CHROMagar, which may actually aid in visualizing the smaller colonies inside the zone of inhibition, resulting in a smaller zone size.
Reading the zone of inhibition around the fluconazole disk is at times difficult and quite subjective. With yeast, the zone of inhibition is read where there is a predominant reduction of growth, unlike the conventional Kirby-Bauer method used for most bacterium-antimicrobial combinations, where the zone of inhibition is distinct (7). Nevertheless, overall, interpretation of results obtained from disk diffusion correlated >80% of the time with the categorical interpretation derived from the reference method. The two very major errors seen with the more resistant isolates of C. glabrata and C. parapsilosis are troubling. However, by extending the incubation time to 48 h, the very major errors seen with these two species were resolved. When using the direct method, extension of the incubation time to 48 h for these two species is recommended. This is also true of the NCCLS M44-A method using MH-GMB, which revealed the same problem when fluconazole zones were read at 24 h with C. glabrata isolates. In fact, the problem was even greater using the method recommended in NCCLS M44-A (7) in comparison to the direct method using CHROMagar described in this report. In summary, use of CHROMagar allowed rapid identification of commonly isolated yeasts from blood cultures and also provided information on the susceptibility to fluconazole to aid in treatment decisions.
REFERENCES
Antoniadou, A., H. A. Torres, R. E. Lewis, J. Thornby, G. P. Bodey, J. P. Tarrand, X. Y. Han, K. V. Rolston, A. Safdar, I. I. Raad, and D. P. Kontoyiannis. 2003. Candidemia in a tertiary care cancer center: in vitro susceptibility and its association with outcome of initial antifungal therapy. Medicine 82:309-321.
Espinel-Ingroff, A., M. Pfaller, S. A. Medsser, C. C. Knapp, S. Killian, H. A. Norris, and M. A. Ghannoum. 1999. Multicenter comparison of the Sensititre YeastOne colorimetric antifungal panel with the National Committee for Clinical Laboratory Standards M27-A reference method for testing clinical isolates of common and emerging Candida spp., Cryptococcus spp., and other yeasts and yeast-like organisms. J. Clin. Microbiol. 37:591-595.
Hazen, K., and S. A. Howell. 2003. Candida, Cryptococcus, and other yeasts of medical importance, p. 1693-1711. In P. M. Murray, E. J. Baron, J. H. Jorgensen, M. A. Pfaller, and R. H. Yolken (ed.), Manual of clinical microbiology, 8th ed. ASM Press, Washington, D.C.
Horvath, L. L., D. R. Hospenthal, C. K. Murray, and D. P. Dooley. 2003. Direct isolation of Candida spp. from blood cultures on the chromogenic medium CHROMagar Candida. J. Clin. Microbiol. 41:2629-2632.
Hospenthal, D. R., C. K. Murray, and M. G. Rinaldi. 2004. The role of antifungal susceptibility testing in the therapy of candidiasis. Diagn. Microbiol. Infect. Dis. 48:153-160.
National Committee for Clinical Laboratory Standards. 2002. Reference method for broth dilution antifungal susceptibility testing of yeasts. Approved standard M27-A2. National Committee for Clinical Laboratory Standards, Wayne, Pa.
National Committee for Clinical Laboratory Standards. 2004. Method for antifungal disk diffusion susceptibility testing of yeasts: proposed guideline M44-A. National Committee for Clinical Laboratory Standards, Wayne, Pa.
Pfaller, M. A., D. J. Diekema, L. Boyken, S. A. Messer, S. Tendolkar, and R. J. Hollis. 2003. Evaluation of the Etest and disk diffusion methods for determining susceptibilities of 235 bloodstream isolates of Candida glabrata to fluconazole and voriconazole. J. Clin. Microbiol. 41:1875-1880.(Grace L. Tan and Ellena M)