Comparison of the Quantum II, API Yeast Ident, and AutoMicrobic systems for identification of clinical yeast isolates

An Update on the Roles of Non-albicans Candida Species in Vulvovaginitis

Candida species are one of the commonest causes of vaginitis in healthy women of reproductive age. Vulvovaginal candidiasis (VVC) is characterized by vulvovaginal itching, redness and discharge. Candida albicans, which is a common genito-urinary tract commensal, has been the prominent species and remains the most common fungal agent isolated from clinical samples of patients diagnosed with VVC. In recent times, however, there has been a notable shift in the etiology of candidiasis with non-albicans Candida (NAC) species gaining prominence. The NAC species now account for approximately 10% to as high as 45% of VVC cases in some studies. This is associated with treatment challenges and a slightly different clinical picture. NAC species vaginitis is milder in presentation, often occur in patients with underlying chronic medical conditions and symptoms tend to be more recurrent or chronic compared with C. albicans vaginitis. C. glabrata is the most common cause of NAC-VVC. C. tropicalis, C. krusei, C. parapsilosis, and C. guilliermondii are the other commonly implicated species. Treatment failure is common in NAC-VVC, since some of these species are intrinsically resistant or show low susceptibilities to commonly used antifungal agents. This article reviews the etiology, pathogenesis, clinical features, diagnosis, and management of NAC vulvovaginitis.

Fungal diagnostics: review of commercially available methods

Fungi and yeasts are critical causes of acute infection. As such, the detection and identification of these organisms are crucial in the diagnosis of affected patient populations. There is a vast array of commercial tests currently available for diagnostic purposes. These vary from traditional culture and biochemical methods to advanced multiparameter molecular tests. Recent technological advances have driven the development of rapid tests which are complementing and in some cases replacing the more traditional methods of detection. Irrespective of the method used the ultimate goal is timely detection of the infectious agent allowing appropriate treatment and improved outcome for the patient.

Enzymatic substrates in microbiology

Enzymatic substrates are powerful tools in biochemistry. They are widely used in microbiology to study metabolic pathways, to monitor metabolism and to detect, enumerate and identify microorganisms. Synthetic enzymatic substrates have been customized for various microbial assays, to detect an expanding range of both new enzymatic activities and target microorganisms. Recent developments in synthetic enzymatic substrates with new spectral, chemical and biochemical properties allow improved detection, enumeration and identification of food-borne microorganisms, clinical pathogens and multi-resistant bacteria in various sample types. In the past 20 years, the range of synthetic enzymatic substrates used in microbiology has been markedly extended supporting the development of new multi-test systems (e.g., Microscan, Vitek 2, Phoenix) and chromogenic culture media. The use of such substrates enables an improvement in time to detection and specificity over conventional tests that employ natural substrates. In the era of intense developments in molecular biology, phenotypic tests involving enzymatic substrates remain useful to analyse both simple and complex samples. Such tests are applicable to diagnostic and research laboratories all over the world.

Can the diagnosis of recurrent vulvovaginal candidosis be improved by use of vaginal lavage samples and cultures on chromogenic agar?

OBJECTIVE

To investigate if introital and vaginal flushing samples inoculated on chromogenic agar could increase the recovery rate and rapid identification of Candida and non-albicans species, as compared to culture of posterior vaginal fornix samples on Sabouraud agar and speciation of isolates by biochemical tests.

METHODS

Samples from the introitus and the posterior vaginal fornix and vaginal lavage samples were collected from 91 women with a history suggestive of recurrent vulvovaginal candidosis (RVVC), and with a suspected new attack of the condition. The specimens were cultured on Sabouraud and CHROMagar. Speciation of yeast isolates was made on the chromogenic agar by API 32C kits and by an atomized system (Vitek).

RESULTS

Forty-six (51%) women were positive for Candida from one or more of the samples. The introital cultures were positive in 43 (47%) women, both on Sabouraud and chromogenic agar. From the posterior vaginal fomix, 42 (46%) women were positive on the Sabouraud and 43 (47%) on chromogenic agar cultures, while the vaginal lavage cultures yielded Candida on those two media in 40 (44%) and 41 (45%) cases, respectively. Candida albicans was the most frequent species recovered, from 40 (87%) cases, followed by C. krusei in 4 (9%), C. glabrata in 2 (4%), and C. parapsilosis in one case. There was only one woman who had a mixed yeast infection, by C. albicans and C. krusei. There was only one discrepancy in the speciation as demonstrated by mean of chromogenic agar and API 32C kit.

CONCLUSIONS

Neither cultures of introital nor of vaginal lavage samples increases the detection rate of Candida in RVVC cases as compared to cultures of posterior vaginal fornix samples. Use of chromogenic agar is a convenient and reliable means to detect colonization by Candida and differentiate between C. albicans and non-albicans species.

Yeast identification in the clinical microbiology laboratory: phenotypical methods

Emerging yeast pathogens are favoured by increasing numbers of immunocompromised patients and by certain current medical practices. These yeasts differ in their antifungal drug susceptibilities, and rapid species identification is imperative. A large variety of methods have been developed with the aim of facilitating rapid, accurate yeast identification. Significant recent commercial introductions have included species-specific direct enzymatic colour tests, differential chromogenic isolation plates, direct immunological tests, and enhanced manual and automated biochemical and enzymatic panels. Chromogenic isolation media demonstrate better detection rates of yeasts in mixed cultures than traditional media, and allow the direct identification of Candida albicans by means of colony colour. Comparative evaluation of rapid methods for C. albicans identification, including the germ tube test, shows that chromogenic media may be economically advantageous. Accurate tests for single species include the Bichrolatex Albicans and Krusei Color tests, both immunologically based, as well as the Remel Rapid Trehalose Assimilation Broth for C. glabrata. Among broad-spectrum tests, the RapID Yeast Plus system gives same-day identification of clinical yeasts, but performance depends on inoculum density and geographic isolate source. The API 20 C AUX system is considered a reference method, but newer systems such as Auxacolor and Fungichrom are as accurate and are more convenient. Among automated systems, the ID 32 C strip, the Vitek Yeast Biochemical Card and the Vitek 2 ID-YST system correctly identify >93% of common yeasts, but the ID-YST is the most accurate with uncommon yeasts, including C. dubliniensis. Spectroscopic methods such as Fourier transformed-infrared spectroscopy offer potential advantages for the future. Overall, the advantages of rapid yeast identification methods include relative simplicity and low cost. For all rapid methods, meticulous, standardized multicenter comparisons are needed before tests are fully accepted.

Variation in Microbial Identification System accuracy for yeast identification depending on commercial source of Sabouraud dextrose agar

The accuracy of the Microbial Identification System (MIS; MIDI, Inc. ) for identification of yeasts to the species level was compared by using 438 isolates grown on prepoured BBL Sabouraud dextrose agar (SDA) and prepoured Remel SDA. Correct identification was observed for 326 (74%) of the yeasts cultured on BBL SDA versus only 214 (49%) of yeasts grown on Remel SDA (P < 0.001). The commercial source of the SDA used in the MIS procedure significantly influences the system's accuracy.

Comparison of three commercial systems for the identification of germ-tube negative yeast species isolated from clinical specimens

Three commercial systems were evaluated for their ability to identify 171 germ-tube negative yeasts isolated from clinical specimens. The Yeast Biochemical Card and Analytical Profile Index 20 AUX identified 97% of 171 strains tested. The Biolog system had poor clinical utility: only 48% of strains were identified. For Yeast Biochemical Card and Analytical Profile Index 20 AUX, 9% and 6%, respectively, required repeat testing and both systems required supplemental tests for 28% of the strains. These observations indicate that considerable expertise and a battery of reagents in addition to the basic systems are required for accurate identification of germ-tube negative yeasts.

Evaluation of six commercial systems for identification of medically important yeasts

Six commercially available systems for the identification of yeasts were evaluated using 133 clinical isolates and four reference strains that had been previously identified by conventional methods and 19 recent clinical isolates that had been identified by the ID32C system (bioMérieux, France). The total identification rates (TIR) established for the total number of strains tested and the database identification rates (DBIR) established for the strains included in the respective manufacturer databases were both determined. After incubation for 4 h, the TIR and DBIR were 78% and 84%, respectively, for the RapID Yeast Plus system (Innovative Diagnostic Systems, USA). After incubation for 24 h, the TIR and DBIR were 32% and 32%, respectively, for the ID32C, 65% and 67% for the Auxacolor system (Sanofi Diagnostics Pasteur, France), 62% and 65% for the Fungichrom I system (International Microbio, France), 52% and 65% for the Fungifast I twin system (International Microbio), and 62% and 68% for the API Candida system (bioMérieux). The maximum TIR and DBIR (+/- 1%) obtained after incubation for 48 h were 86% and 88% for the Auxacolor, 85% and 89% for the Fungichrom I, 78% and 98% for the Fungifast I twin, and 82% and 91% for the API Candida. For the ID32C, the maximum TIR and DBIR were 98% and 98%, respectively, but these values were obtained only after 72 h of incubation. In addition, the six systems varied in their ease of use and readings. In conclusion, based on results obtained with 156 strains, the Auxacolor and Fungichrom systems seem the most appropriate for use in a clinical microbiology laboratory, due to their ease of use and reading, their rapidity, their cost per test, and their relatively high TIR results, which indicated acceptable performance with strains frequently isolated in our hospital. For a reference identification, the ID32C remains the sole system usable.

Rapid identification of Candida albicans and other human pathogenic yeasts by using short oligonucleotides in a PCR

A PCR system that can quickly and accurately identify 14 species of human pathogenic yeasts was developed. The procedure distinguished between nine species of a closely related clade, Lodderomyces elongisporus, Candida parapsilosis, a new Candida sp., C. sojae, C. tropicalis, C. maltosa, C. viswanathii, C. albicans, and C. dubliniensis and between another five more divergent species, Pichia guilliermondii, C. glabrata, C. zeylanoides, C. haemulonii, and C. haemulonii type II. A rapid DNA extraction procedure that yields purified DNA in about 1 h is also described. The system uses uniform conditions with four primers for each reaction, two 40- to 50-mer universal primers that serve as a positive control and two 23- to 30-mer species-specific primers. Species-specific primers were derived from a 600-nucleotide variable region (D1/D2) at the 5' end of the large-subunit (26S) ribosomal DNA gene and were generally designed to use mismatches at the 3' end. Universal primers were developed from conserved nucleotide sequences in the small-subunit (18S) rRNA gene. In this system, a control 1,200- to 1,300-base DNA fragment was produced in all reactions and a smaller 114- to 336-base DNA fragment was produced if the chromosomal DNA from the target species was present. The PCR procedure is rapid and easy to interpret and may be used with mixed cultures.

Limitations of the current microbial identification system for identification of clinical yeast isolates

The ability of the rapid, computerized Microbial Identification System (MIS; Microbial ID, Inc.) to identify a variety of clinical isolates of yeast species was compared to the abilities of a combination of tests including the Yeast Biochemical Card (bioMerieux Vitek), determination of microscopic morphology on cornmeal agar with Tween 80, and when necessary, conventional biochemical tests and/or the API 20C Aux system (bioMerieux Vitek) to identify the same yeast isolates. The MIS chromatographically analyzes cellular fatty acids and compares the results with the fatty acid profiles in its database. Yeast isolates were subcultured onto Sabouraud dextrose agar and were incubated at 28 degrees C for 24 h. The resulting colonies were saponified, methylated, extracted, and chromatographically analyzed (by version 3.8 of the MIS YSTCLN database) according to the manufacturer's instructions. Of 477 isolates of 23 species tested, 448 (94%) were given species names by the MIS and 29 (6%) were unidentified (specified as "no match" by the MIS). Of the 448 isolates given names by the MIS, only 335 (75%) of the identifications were correct to the species level. While the MIS correctly identified only 102 (82%) of 124 isolates of Candida glabrata, the predictive value of an MIS identification of unknown isolates as C. glabrata was 100% (102 of 102) because no isolates of other species were misidentified as C. glabrata. In contrast, while the MIS correctly identified 100% (15 of 15) of the isolates of Saccharomyces cerevisiae, the predictive value of an MIS identification of unknown isolates as S. cerevisiae was only 47% (15 of 32), because 17 isolates of C. glabrata were misidentified as S. cerevisiae. The low predictive values for accuracy associated with MIS identifications for most of the remaining yeast species indicate that the procedure and/or database for the system need to be improved.

[Candida identification: experiences with the Vitek automated system]

We report about our experiences obtained with the Vitek system (bioMerieux, Nürtingen) for identification of 1160 clinical Candida isolates. The Vitek system correctly identified 1005 (86.6%) Candida isolates (probability > or = 85%): 776 (66.9%) strains were identified after 24 h of incubation, 229 (19.7%) required 48 h of incubation. In 4.4% (n = 51) we found results with a lower probability (< 85%). After 24 h of incubation 38 (14.6%) of the 260 C. tropicalis isolates tested were misidentified as C. parapsilosis; we obtained the correct results after 48 h of incubation. 104 (9%) isolates were wrongly identified after 48 h of incubation. The majority of these strains (n = 82) were biochemically minimally reactive Candida species. Overall, the Yeast Biochemical Card system provides a reliable method for the rapid, automated identification of medically important Candida species.

Species-specific identification of Candida krusei by hybridization with the CkF1,2 DNA probe

The species specificity of the Candida krusei DNA fingerprinting probe CkF1,2 has been investigated. A total of 149 pathogenic and nonpathogenic fungal and bacterial DNAs were screened with CkF1,2. The probe was cold labeled with peroxidase, and its specificity was assessed by using Southern blot, dot blot, and colony blot hybridization. Its sensitivity was determined by dot blot hybridization. The CkF1,2 probe proved to be species specific. It hybridized with DNA for the 112 C. krusei strains studied, whereas it failed to hybridize under low-stringency conditions to 37 DNAs from 27 different yeast species, including Candida albicans, Candida glabrata, Candida norvegensis, Candida inconspicua, Candida tropicalis, Candida valida, Candida zeylanoides, and Yarrowia lipolytica, as well as DNAs from the filamentous fungi and bacteria tested. However, CkF1,2 hybridized strongly with DNA of the yeast species Issatchenkia orientalis, the putative ascogenous perfect state of C. krusei. Amounts as small as 60 to 120 ng of C. krusei target DNA were detected by dot blot hybridization with CkF1,2. It permitted the direct screening of colony blots for early identification. The CkF1,2 probe has potential value as a diagnostic reagent for identifying C. krusei.

Comparison of the new API Candida system to the ID 32C system for identification of clinically important yeast species

API Candida was evaluated in comparison with the ID 32C system for the identification of 619 yeast isolates. The sensitivity of API Candida for the identification of the 15 species it claims to identify with and without additional tests was 97.4% (593 of 609) and 75.2% (458 of 609), respectively. The API Candida system is easy to use and rapid (result in 18 to 24 h).

Application of CHROMagar Candida for rapid screening of clinical specimens for Candida albicans, Candida tropicalis, Candida krusei, and Candida (Torulopsis) glabrata

CHROMagar Candida is a new differential culture medium that allows selective isolation of yeasts and simultaneously identifies colonies of Candida albicans, C. tropicalis, and C. krusei. We evaluated the use of this medium with 316 yeast isolates including 247 isolated directly on CHROMagar from clinical material. Over 95% of stock and clinical isolates of C. albicans, C. tropicalis, and C. krusei were correctly identified on the basis of colony morphology and pigmentation on CHROMagar. Additionally, CHROMagar also allowed the identification of C. (Torulopsis) glabrata at a similar level of accuracy. The overall agreement between two observers in reading the CHROMagar plates was 95%. Growth of Candida sp. isolates on CHROMagar had no adverse effect on antifungal MICs or Vitek identification results. In parallel, cultures of 548 stool and rectal swab specimens set up on CHROMagar and Sabouraud glucose agar (SGA) were positive in 234 instances. CHROMagar was positive and SGA was negative for 11 specimens, and CHROMagar was negative and SGA was positive for 18 specimens. A single yeast species was isolated on both media from 162 specimens, and in 146 (90%) of these specimens the same species was detected on both CHROMagar and SGA. A total of 43 of the 234 positive cultures contained mixtures of yeast species. Twenty (47%) of these mixed cultures were detected only on CHROMagar. CHROMagar is extremely useful in making a rapid presumptive identification of common yeast species. This capability plus the ability to detect mixed cultures of Candida spp. promises to improve and streamline the work flow in the mycology and clinical microbiology laboratory.

Evaluation of the MicroScan enzyme-based system for the identification of foodborne yeasts

Eighty-nine strains representing 36 species of foodborne yeasts isolated from fruit juice concentrates were identified using the Baxter MicroScan enzyme-based kit, conventional tests according to a simplified identification method (SIM), and the API 20C kit. Of the 15 test species included in the MicroScan database, only 40% were correctly identified; 13% gave scores of unacceptably low probabilities, 20% were misidentified, and 27% could not be identified. Of the 21 test species not in the MicroScan database, 38% were misidentified and 62% produced biocodes with between-species differences not larger than differences between strains within species. The reliability of the MicroScan enzyme-based system is questioned, in that different results were sometimes obtained upon retesting the same strains. The MicroScan enzyme-based system is rapid, providing results within 4 h. However, because of its restricted and specific database and unreliability, the system appears to be unsuited for the identification of foodborne yeasts.

Evaluation of the AUXACOLOR system, a new method of clinical yeast identification

AIMS

To compare the AUXACOLOR yeast identification system with the API 20C system.

METHODS

Yeast isolates (n = 215), comprising 16 species, were identified using the AUXACOLOR system and the API 20C system. Isolates that could not be identified with the API 20C system or which produced discrepant results in the two systems were identified by assimilation and fermentation procedures.

RESULTS

AUXACOLOR correctly identified 150 (85.7%) of 175 germ tube negative isolates while API 20C identified 155 (88.6%). Incorrect identifications were more common with API 20C (7.4%) than with AUXACOLOR (3.7%). Of 110 isolates of four common pathogens (Candida glabrata, C parapsilosis, C tropicalis, and Cryptococcus neoformans), 82.7% (91/110) were identified by AUXACOLOR while API 20C identified 74.5% (82/110). Of 65 less common germ tube negative isolates, 55.4% (36/65) were identified by AUXACOLOR while API 20C identified 63.1% (41/65).

CONCLUSION

Although it has a limited database of 26 species, the AUXACOLOR system is a useful method for identification of germ tube negative clinical yeast isolates. Compared with the API 20C, the AUXACOLOR system is simpler and quicker to set up, easier to interpret, and comparable in cost.

Misidentification of clinical yeast isolates by using the updated Vitek Yeast Biochemical Card

The Vitek Yeast Biochemical Card (YBC) is widely used as a rapid identification (RI) (within 48 h) system for clinical yeast isolates. We compared the RI results obtained by the YBC technique with matched results obtained with the API 20C system. The RI of germ tube-negative yeasts isolated from 222 clinical specimens was performed with the YBC system, and the results were compared with those of standard identifications obtained by using the API 20C system and morphology, with additional biochemical reactions performed as required. Commonly isolated yeasts (Candida albicans [n = 29], Candida tropicalis [n = 40], Torulopsis [Candida] glabrata [n = 28], Candida parapsilosis [n = 12], and Cryptococcus neoformans [n = 14]) were generally well identified (115 of 123 [93%] identified correctly, with only C. albicans, C. tropicalis, and C. neoformans mis- or unidentified more than once). The RI of less commonly isolated yeasts included in the YBC database, however, was less successful (54 of 99 [55%] correct). The YBC card failed to identify 42% (10 of 24) of Candida krusei isolates, 80% (4 of 5) of Candida lambica isolates, 88% (7 of 8) of Trichosporon beigelii isolates, and 83% (10 of 12) of Cryptococcus isolates (non-C. neoformans species). For most identification failures (79%; 42 of 53) there was no identification by the end of 48 h; the other identification failures (21%; 11 of 53) gave definite but incorrect identifications. Of eight rare clinical yeast isolates not included in the Vitek database, six were correctly, not identified, while two (25%) were falsely assigned a definite RI (one Hansenula fabianii isolate was identified as Rhodotorula glutinis, and one Hansenula isolate [non-Hansenula anomala] was identified as Hansenula anomala). While the Vitek YBC rapidly and adequately identifies common yeast isolates, it fails in the RI of more unusual organisms.

Comparison of updated Vitek Yeast Biochemical Card and API 20C yeast identification systems

The updated Vitek Yeast Biochemical Card (YBC) was compared with the API 20C by using 409 germ tube-negative yeasts and Geotrichum spp. that were either clinical or proficiency sample isolates. The API 20C was the reference standard. The 409 isolates represented nine genera and 21 species. Morphology agars were inoculated and interpreted for each isolate. The API 20C identified 406 isolates (99.3%), while the Vitek YBC identified 367 (89.7%). Both systems identified the majority of yeasts after 24 h of incubation--73.4% were identified by the API 20C and 77.4% were identified by the Vitek YBC. The Vitek 24-h reading had some incorrect identifications. These included 14 isolates of Candida tropicalis that were identified as Candida parapsilosis (91 to 97% reliability) and 3 isolates of Candida krusei that were called Blastoschizomyces capitatus (Geotrichum capitatum), Candida rugosa, and Candida zeylanoides. In total, the Vitek YBC misidentified 30 isolates, while the API 20C misidentified 3 isolates. In addition, results for 14 isolates with the Vitek YBC were listed under the category "no identification." Morphology agars were required for identification with 89 isolates (21.9%) when the API 20C was used and with 50 isolates (12.6%) when the Vitek YBC was used. Apart from the price of the Vitek instrument, the API 20C costs $1.28 more per test than the Vitek YBC. Overall, the updated Vitek YBC compares favorably with the API 20C in the identification of common yeasts such as Torulopsis glabrata, C. parapsilosis, and Cryptococcus neoformans. However, problems were encountered with the Vitek system in the identification of C. tropicalis, C. krusei, Trichosporon spp., and some Cryptococcus spp. The routine use of morphology agars with either method is recommended.

Evaluation of two commercialized systems for the rapid identification of medically important yeasts

A total of 77 recent clinical isolates of Candida albicans and other medically important yeasts were identified by two different commercial tests, Rapidec albicans (API-bioMérieux) and Fongiscreen 4H (Sanofi Diagnostics Pasteur), and conventional mycological methods. The strains were from 13 different species of yeasts and consisted of strains of 36 C. albicans, three of Candida famata, nine of Candida (Torulopsis) glabrata, five of Candida guilliermondii, two of Candida kefyr, three of Candida krusei, one of Candida lusitaniae, four of Cryptococcus neoformans, five of Candida parapsilosis, six of Candida tropicalis, one of Candida viswanathii, one of Rhodotorula rubra and one of Saccharomyces cerevisiae. According to the reactivity profiles of the isolates, identification was always correct with Fongiscreen 4H and was correct in 97.3% of the strains with Rapidec albicans. The latter test did not identify two C. albicans isolates that were correctly identified by Fongiscreen 4H. Both methods (97.3% correlation) were very useful for identification of C. albicans achieving the aim of their manufacturers. Additionally, Fongiscreen 4H was very useful for the identification of three other species of yeasts: C. glabrata, C. tropicalis and Cr. neoformans. The results of our study indicate that the accuracy of Rapidec albicans and Fongiscreen 4H is similar to that of the conventional methods used in this study for the identification of C. albicans. The same is true of Fongiscreen 4H in the identification of C. glabrata, C. tropicalis and Cr. neoformans. Both tests could be rapid and easy-to-perform tools in the clinical microbiology laboratory, but differences in cost must be taken into account.

Simplified techniques for identifying foodborne yeasts

Four problematic areas associated with the identification of foodborne yeasts are discussed. These consist of (1) the inability of conventional identification tests to recognize some common and important foodborne yeasts characterized by genomic differences (e.g., Saccharomyces cerevisiae, S. bayanus and S. pastorianus); (2) the delay in application of non-traditional identification methods such as DNA fingerprinting, chromosome karyotyping, protein electrophoretic patterns and fatty acid profiles for routine identification purposes; (3) the lack of commercially available manual or automated identification systems dedicated to the diagnosis of foodborne yeasts; and (4) the disregard for considering ecological frequency of yeasts in computerized probabilistic identification systems.

Evaluation of Gen-Probe's Histoplasma capsulatum and Cryptococcus neoformans AccuProbes

Gen-Probe's DNA probes were evaluated for use in the identification of clinical isolates of Histoplasma capsulatum var. capsulatum and Cryptococcus neoformans. Ninety-five mould-phase fungi were probed, including 41 isolates of H. capsulatum var. capsulatum. Similarly, 98 yeasts, including 42 C. neoformans isolates, were examined by using the C. neoformans DNA probe. In the study, both probes demonstrated 100% specificity and 100% sensitivity. Their use in the clinical laboratory may significantly reduce the time required for definitive identification of fungi.

Automated systems for identification of microorganisms

Automated instruments for the identification of microorganisms were introduced into clinical microbiology laboratories in the 1970s. During the past two decades, the capabilities and performance characteristics of automated identification systems have steadily progressed and improved. This article explores the development of the various automated identification systems available in the United States and reviews their performance for identification of microorganisms. Observations regarding deficiencies and suggested improvements for these systems are provided.

Evaluation of a new method for identification of Cryptococcus neoformans which uses serologic tests aided by selected biological tests

A new method for identifying Cryptococcus neoformans isolates and their serotypes by the slide agglutination test using five kinds of factor sera, with the aid of nitrate reduction, phenol oxidase, and growth at 37 degrees C tests was evaluated by using 36 reference strains and 75 clinical isolates of C. neoformans. The results showed that the reference strains were identified exactly as they were labeled, and clinical isolates were identified as C. neoformans serotypes A, D, and AD. C. neoformans could be distinguished from other Cryptococcus species that cross-reacted with factor sera by their ability to grow at 37 degrees C. These results indicate that the slide agglutination test combined the use of factor sera for isolates which grow at 37 degrees C is a useful method for identification of C. neoformans and their serotypes and that the nitrate reduction test (negative in 100% of the isolates) and the phenol oxidase test (positive in approximately 95% of the isolates) can be used to confirm that the species is C. neoformans.

Evaluation of the MicroScan Rapid Yeast Identification panel

The MicroScan Rapid Yeast Identification (RYI) panel is a 4-h microdilution system for identification of clinical yeastlike isolates. Its accuracy was evaluated by using 357 isolates encompassing 11 genera and 30 species. The RYI panel identifications were compared with those obtained by the API 20C system assisted with morphological characterization on cornmeal-Tween 80 agar. The panels were read both visually and with the AutoScan-4, a computer-controlled microplate reader. Both the RYI panel and the API 20C system correctly identified 78% of the strains within 4 and 72 h, respectively, with no additional tests. Supplementary tests recommended by the manufacturers made it possible to identify up to 96.6% (AutoScan-4) and 98.9% (API 20C) of the strains. The accuracy of the RYI panel was 99.5% with common strains and 92.1% with less common strains. The RYI panel misidentified 10 or 12 strains and failed to identify 2 or 3 strains, depending on whether it was read with the AutoScan-4 or visually. Errors occurred with one strain of Torulopsis glabrata and the less common yeasts T. candida, Candida lusitaniae, C. lambica, C. rugosa, C. stellatoidea, Cryptococcus albidus, C. laurentii, and C. uniguttulatus. Overall, the RYI panel appears to be a reliable system for identification of the more common clinical yeast isolates.

Comparison of the autoSCAN-W/A rapid bacterial identification system and the Vitek AutoMicrobic system for identification of gram-negative bacilli

The autoSCAN-W/A (W/A; Baxter MicroScan, West Sacramento, Calif.) with the new fluorometric Rapid Neg Combo 1 (RNC) panel is a fully automated fluorometric system for identification of both enteric and nonenteric gram-negative bacilli within 2 h. We compared the W/A with the Vitek AutoMicrobic System (Vitek AMS; Vitek Systems, Inc., Hazelwood, Mo.) for identification of 383 clinical isolates of gram-negative bacilli. The API 20E (Analytab Products, Plainview, N.Y.) and conventional biochemical testing were used as the reference systems. The W/A correctly identified 336 isolates (87.7%) to the species level and classified an additional 29 isolates (7.6%) as correct with low probability (overall identification = 95.3%); the Vitek AMS correctly identified 355 isolates (92.7%) to the species level and classified an additional 8 isolates (2.1%) as correct with low probability (overall identification = 94.8%). A common set of 134 isolates of gram-negative bacilli was tested in both participating laboratories as a means of assessing interlaboratory agreement with both the W/A and the Vitek AMS. The overall agreements between the two laboratories were 86% with the W/A and 92% with the Vitek AMS. The W/A performed comparably to the Vitek AMS for identification of most gram-negative bacilli, actually exceeding the Vitek AMS for identification of nonenteric bacilli. Rapid time to identification and a high level of automation make the W/A an attractive system for clinical microbiology laboratories.

Evaluation of the Baxter-MicroScan 4-hour enzyme-based yeast identification system

A new 4-h Yeast Identification Panel (YIP; Baxter-MicroScan, W. Sacramento, Calif.) was compared with the API 20C Yeast Identification System (Analytab Products, Inc., Plainview, N.Y.) in the identification of recent clinical yeast isolates. The YIP had a 94% correlation (288 of 306) in identifying 22 species within the genera Candida, Hansenula, Pichia, Rhodotorula, Saccharomyces, and Torulopsis. Correlation dropped to 65% for those species within the genera of slower growing yeasts, i.e., Blastoschizomyces spp., Crpytococcus spp., Geotrichum spp., Hyphopichia spp., Phaeococcomyces spp., Prototheca spp., and Trichosporon spp. Overall correlation with the API 20C was 92% (365 of 401) for those taxa included in the data base and 85% (373 of 437) for all yeasts encountered in the study. There were 36 (8.2%) discrepant identifications, which were due in part to the limited data base. Expansion of the data base plus the easy inoculation, reading, and rapid results of the YIP should make it an excellent method for yeast identification.

Multicenter evaluation of Microring YT, a new method of yeast identification

Microring YT was evaluated and compared with established methods for the identification of 142 clinical yeast isolates. Only 75 isolates (52.8%) were correctly identified by Microring YT. Results with this test were often difficult to read and subject to interlaboratory variations.

Evaluation of the updated Vitek yeast identification data base

Using 398 isolates of yeasts and yeastlike fungi comprising 9 genera and 26 species, as well as the hyphomycete Geotrichum candidum and the achlorophyllous alga Prototheca wickerhamii, we compared the API 20C yeast identification system with the modified Vitek yeast identification system with an expanded data base. We found 11 discrepancies between the two systems: five (1.3%) of the isolates (Blastoschizomyces capitatus, 1; Candida albicans, 1; Hansenula anomala, 1; Rhodotorula minuta, 2) had biocodes not included in the expanded Vitek data base, and six (1.5%) of the isolates (Candida lusitaniae, 1; Candida parapsilosis, 1; Cryptococcus uniguttulatus, 1; H. anomala, 1; Torulopsis candida, 2) were misidentified by the Vitek system. Overall, the efficacy of the Vitek system compares favorably with that of the API 20C in the identification of clinically important yeasts.

Laboratory identification and sensitivity testing of yeast isolates

Methods used for identification and sensitivity testing of yeasts are presented. Identification depends on both morphologic features and biochemical characteristics. The germ tube test is a simple test for the identification of the commonest yeast pathogen, Candida albicans. Species not identified by this method can be identified by means of several other tests, the most important being the carbohydrate assimilation test. Several kits are commercially available and most appear to be satisfactory. Antifungal susceptibility tests are not standardized, and results are often subject to considerable variation. Until improved methods are developed, it might be advisable that smaller laboratories only do a preliminary screening test for flucytosine susceptibility and, if indicated, also for amphotericin B susceptibility. Important isolates should be sent to laboratories with a special interest in antifungal susceptibility testing for more extensive testing.