![[Feedback]](/icons/banner/feedbackACT.gif)
![[For Subscribers]](/icons/banner/subscriberACT.gif)
![[Archive]](/icons/banner/archiveACT.gif)
![[Search]](/icons/banner/searchACT.gif)
![[Contents]](/icons/banner/tocACT.gif)
Previous Article | Next Article 
Journal of Clinical Microbiology, June 1999, p. 2080-2083, Vol. 37, No. 6
0095-1137/99/$04.00+0
Copyright © 1999, American Society for Microbiology. All rights reserved.
Variation in Microbial Identification System
Accuracy for Yeast Identification Depending on Commercial Source of
Sabouraud Dextrose Agar
James A.
Kellogg,1,*
David A.
Bankert,1 and
Vishnu
Chaturvedi2
Clinical Microbiology Laboratory, York
Hospital, York, Pennsylvania 174051 and
Laboratories for Mycology, Axelrode Institute, New York
State Department of Health, Albany, New York
12208-20022
Received 28 December 1998/Returned for modification 1 March
1999/Accepted 22 March 1999
 |
ABSTRACT |
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.
| |
TEXT |
The Microbial Identification System
(MIS; Microbial ID, Inc., Newark, Del.) is an automated approach to
rapid identification of unknown microbial isolates. The MIS consists of
a gas chromatograph with a flame ionization detector, an autosampler,
an integrator, and a computer. The MIS computer searches a software
library of fatty acid compositions, compares the isolate's fatty acid
profile with those of known species, and reports the most likely
species name along with the extent of correlation (the similarity index [SI]) of the isolate's profile with a species in the database (22).
The MIS has provided an accurate, rapid, and cost-effective alternative
for identification of many aerobic gram-positive and gram-negative
bacterial isolates (2, 12, 16). Two recent reports, however,
have noted that only 70 to 71% of yeast isolates were correctly
identified by the MIS to the species level (6, 13). The
yeast database for the MIS was derived by the system's manufacturer
from yeasts grown on BBL 11584 Sabouraud dextrose agar (SDA), which was
purchased in the dehydrated form (Becton-Dickinson Microbiology
Systems, Cockeysville, Md.) and freshly prepared in house prior to use
by Microbial ID (17). Most clinical laboratories, however,
buy their culture media prepoured from commercial suppliers. Of the two
published studies concerning yeast identification using the MIS, one
used yeasts cultured on prepoured BBL SDA (13). The
commercial source of SDA for the other study (6) is unknown. The present study evaluated the relative ability of the MIS to accurately identify yeasts grown on prepoured SDA plates obtained from
two widely used commercial sources: BBL and Remel, Inc. (Lenexa, Kans.). The intent was to determine if significantly different levels
of accuracy of MIS performance could be experienced when yeasts were
cultured on SDA purchased from different commercial sources.
The majority of yeasts (375; 86%) used during the present study were
freshly isolated from clinical specimens. In addition, 63 (14%) of the
isolates were stock cultures of less frequently recovered yeasts (from
the Laboratories for Mycology, New York State Department of Health),
which had previously been recovered from clinical specimens. More than
75% of the total isolates studied had been recovered from genital,
urine, and wound specimens submitted for culture. Multiple isolates of
the same species from the same patients were excluded from the study.
Isolates were initially subcultured to SDA (BBL) and to cornmeal agar
with 0.5% Tween 80 (CMT); each subculture was incubated at 25°C.
Conventional identification.
Each of the fresh clinical
isolates, as well as the stock cultures of yeasts, was preliminarily
identified by determination of its microscopic morphology on CMT and by
its colony morphology and pigment production on SDA. Clinical isolates
were identified as Candida albicans by their typical
microscopic appearances on CMT, including the production of
chlamydospores (10, 11, 23). Isolates of Candida
dubliniensis were identified by their ability to produce
chlamydospores and their inability to grow at 45°C (19).
Isolates of most of the remaining species were identified by using the
Yeast Biochemical Card (YBC; bioMerieux Vitek, Hazelwood, Mo.)
(7-9, 18), which was inoculated and incubated according to
the manufacturer's specifications as previously described
(13). Fresh clinical isolates which could not be
conclusively identified by using the combination of their microscopic
morphology and the YBC were identified by using the API 20C Aux
(bioMerieux Inc.) (9, 18, 20) and/or additional conventional
tests as appropriate and as previously described (13). Stock
cultures of clinical isolates were identified by the API 20C
(bioMerieux Vitek). The identification carried out as described above
was considered the correct identification.
Chromatographic analysis.
Isolated colonies of each yeast to
be studied were quadrant streaked onto two commercially obtained SDA
plates (BBL and Remel). The cultures were incubated aerobically at
28 ± 1°C for 24 ± 2 h, and extracts were prepared
and analyzed chromatographically as specified in the MIS instructions
(22) and as previously described (13). Version
3.8 of the YSTCLN database in the MIS computer was used to identify
isolates. For each isolate, the computer printout either listed one or
more possible species choices with an SI for each choice ranging from 0 to 1.000, or it reported "no match," the inability of the MIS to
identify the isolate. For the present study, the MIS result was
considered correct if the correct species name of an isolate was listed
on the MIS printout as the first choice, regardless of the SI.
When the MIS result was either a misidentification to the species level
or "no match," the microscopic morphology of the isolate was
determined again and the API 20C Aux or appropriate conventional test
systems were inoculated to confirm the species' identification. In
addition, a fresh extract from a new subculture of the isolate on the
appropriate commercially obtained SDA, incubated at 28°C, was
analyzed in the chromatograph a second time. If an isolate was
misidentified the first time it was analyzed in the MIS, it was counted
as a misidentification, regardless of whether it was correctly or
incorrectly identified by the system when reanalyzed in the MIS. If an
isolate was unidentified by the MIS when first tested and then
correctly or incorrectly identified when the chromatography was
repeated, it was counted as a correct or incorrect identification, respectively. The Z test for differences in proportions for independent samples was used for statistical analysis of results (24). A P value of <0.05 was selected as the minimum level
determining significance.
MIS performance by commercial source of SDA.
Of 438 yeast
isolates from the 22 species of yeasts that were studied, 326 (74%)
versus only 214 (49%) were correctly identified by the MIS to the
species level following growth on BBL and Remel SDA, respectively
(Table 1) (P < 0.001).
Significantly more isolates from BBL SDA were correctly identified by
the MIS to the species level when C. albicans (P < 0.01), Candida glabrata (P < 0.001), Candida guilliermondii (P < 0.05), Candida tropicalis (P < 0.001) and Saccharomyces cerevisiae (P < 0.005)
were chromatographically analyzed. Of the 438 yeasts studied, 68 (16%)
and 100 (23%) were misidentified by the MIS when the yeasts were
cultured on BBL and Remel SDA, respectively (Table
2) (P < 0.01). When
cultured on Remel SDA, 14 (67%) of the 21 misidentified isolates of
C. albicans were called C. tropicalis and 37 (100%) of the misidentified isolates of C. glabrata were
named S. cerevisiae by the MIS. In all, 44 (10%) versus 124 (28%) of the yeasts cultured on BBL and Remel SDA, respectively, were
unidentified by the MIS (P < 0.001).
View this table:
[in this window]
[in a new window]
|
TABLE 1.
Comparison of the Microbial Identification System with
conventional tests for identification of yeasts grown on premade SDA
plates from BBL and Remel
|
|
View this table:
[in this window]
[in a new window]
|
TABLE 2.
Misidentifications of yeast species by the Microbial
Identification System based on commercial source of SDA medium
|
|
Only 24 isolates (5 Candida dubliniensis, 4 Candida
famata, 2 Crytococcus albidus, 2 Cryptococcus
humicolus, 4 Cryptococcus terreus, 6 Hansenula
anomala, and 1 Rhodotorula pilamanae) for which the MIS
software library had no data were studied. Of these isolates, 20 (83%)
and 18 (75%) were misidentified by the MIS when cultures were grown on
BBL and Remel SDA, respectively (a difference that was not significant
[NS]), and the remaining isolates were unidentified by the system. Of
the 438 yeast isolates studied, 138 (32%) cultured on BBL SDA and 246 (56%) grown on Remel SDA had to be chromatographically analyzed a
second time (P < 0.001) because the initial results
were either unidentified ("no match") or incorrect. For these
reanalyzed isolates, results from only 26 (19%) of the 138 BBL
cultures and 22 (9%) of the 246 Remel cultures changed from "no
match" when first tested to correct. Another 20 (14%) of the yeasts
retested from BBL SDA and 17 (7%) of those retested from Remel SDA
changed from incorrect when first analyzed chromatographically to
correct. All remaining yeasts that had to be chromatographically
analyzed a second time gave results that were either incorrect or "no match."
More than 25 production lots of SDA from each manufacturer were used
during the present study. No significant differences were observed in
MIS performance with yeasts cultured on different production lots of
SDA from either manufacturer. While the pH (5.6) was the same from each
commercial source, the ingredients were similar but not identical
(Table 3) (1, 21). When
different species calls were made by the MIS for the same isolate grown on the two commercial sources of SDA, the different sources of metabolic substrates used in the two culture media resulted in qualitative or quantitative differences in the isolate's fatty acid
profile and in an MIS misidentification.
Since the conclusion of the present study, the authors have compared
MIS identification of 91 yeast isolates (14 species) from both freshly
prepared BBL 11584 SDA and prepoured BBL SDA (13a). Of these
isolates, the MIS correctly identified 72 (79%) from prepoured SDA
versus 65 (71%) from freshly prepared SDA (NS). MIS results obtained
with the prepoured medium appear to be at least as accurate as those
obtained with the fresh product.
The results of the present study indicate that accurate MIS
identification of yeasts to the species level is dependent, at least in
part, on the commercial source of prepoured SDA medium that may be used
to culture the organisms just prior to chromatographic analysis. The
percentage of yeasts correctly identified with BBL SDA in the present
study (74%) was similar to both our previous findings with BBL SDA
(70% [13]) and the findings of another study (71%
[6]) using SDA from an unspecified source, but it was
significantly greater than that obtained when yeasts were cultured
during the present study on SDA obtained from Remel (49%).
The percentage of yeasts correctly identified in this or other studies
will also be influenced, in part, by the number and variety of each
species included in the study. For example, the MIS correctly
identified (from BBL SDA) more than 90% of the isolates of C. guilliermondii, Candida parapsilosis,
Rhodotorula rubra, and S. cerevisiae (Table
1). In addition, while the MIS correctly identified 88% of the
isolates of C. glabrata which had been cultured on BBL SDA,
only 6% of the isolates of that species were correctly identified when
cultured on Remel SDA. A similarly low level of correct species
identification (33%) was observed for S. cerevisiae isolates cultured on Remel SDA (while the remaining 67% were
unidentified), compared to the accuracy (100%) for isolates of that
species grown on BBL SDA. Therefore, if a larger number of isolates of
yeast species such as C. glabrata and S. cerevisiae had been included in the study, the overall percentage
of yeasts accurately identified by the MIS, at least when cultured on
Remel SDA, would have been reduced.
In the present study, yeast isolates were frequently misidentified as
other species that have traditionally been either more susceptible or
more resistant to antifungal therapy. Misidentification of yeast
species may lead to the selection of inappropriate therapy by
physicians, both because many yeast species respond differently to
amphotericin B and the azole agents (3-5, 11, 14, 15, 23,
25) and because the majority of clinical laboratories in the
United States usually report yeast species identifications without
accompanying antifungal drug susceptibility results. In the present
study, for example, of the isolates of C. albicans (which is
usually susceptible to amphotericin B and the azoles) studied, 3 (4%)
were misidentified by the MIS as C. guilliermondii (which
may be resistant to amphotericin B and fluconazole [11, 15,
23]) when cultured on either BBL or Remel SDA and 14 (20%) were misidentified as C. tropicalis (which may be resistant
to both amphotericin B and the newer azoles [11, 15,
23]) when grown on Remel SDA prior to chromatographic analysis
(Table 2).
As previously suggested (13, 23), the database for the
potentially useful MIS yeast identification system needs to be improved. The database should be derived from yeasts cultured on
commercially available, prepoured SDA because that, not medium prepared
in house, is what is used most commonly and most practically in
clinical laboratories.
| |
ACKNOWLEDGMENTS |
Statistical analysis of the results was performed by Sally
Cavanaugh, York Hospital Department of Research.
| |
FOOTNOTES |
*
Corresponding author. Mailing address: Clinical
Microbiology Laboratory, York Hospital, 1001 S. George St., York,
PA 17405. Phone: (717) 851-2393. Fax: (717) 851-2707. E-mail:
jkellogg{at}yorkhospital.edu.
| |
REFERENCES |
| 1.
|
Becton Dickinson Microbiology Systems.
1997.
Quality control/product information. Sabouraud dextrose agar. P21180.
Becton Dickinson Microbiology Systems, Cockeysville, Md.
|
| 2.
|
Birnbaum, D.,
L. Herwaldt,
D. E. Low,
M. Noble,
M. Pfaller,
R. Sherertz, and A. W. Chow.
1994.
Efficacy of Microbial Identification System for epidemiologic typing of coagulase-negative staphylococci.
J. Clin. Microbiol.
32:2113-2119[Abstract/Free Full Text].
|
| 3.
|
Blinkhorn, R. J.,
D. Aldelstein, and P. J. Spagnuolo.
1989.
Emergence of a new opportunistic pathogen, Candida lusitaniae.
J. Clin. Microbiol.
27:236-240[Abstract/Free Full Text].
|
| 4.
|
Body, G. P.
1992.
Azole antifungal agents.
Clin. Infect. Dis.
14(Suppl.):161-169.
|
| 5.
|
Christenson, J. C.,
A. Guruswamy,
G. Mukwaya, and P. J. Rettig.
1987.
Candida lusitaniae: an emerging human pathogen.
Pediatr. Infect. Dis. J.
6:755-757[Medline].
|
| 6.
|
Crist, A. E., Jr.,
L. M. Johnson, and P. J. Burke.
1996.
Evaluation of the Microbial Identification System for the identification of clinically isolated yeasts.
J. Clin. Microbiol.
34:2408-2410[Abstract].
|
| 7.
|
Dooley, D. P.,
M. L. Beckius, and B. S. Jeffrey.
1994.
Misidentification of clinical yeast isolates by using the updated Vitek Yeast Biochemical Card.
J. Clin. Microbiol.
32:2889-2892[Abstract/Free Full Text].
|
| 8.
|
El-Zaatari, M.,
L. Pasarell,
M. R. McGinnis,
J. Buckner,
G. A. Land, and I. F. Salkin.
1990.
Evaluation of the updated Vitek yeast identification database.
J. Clin. Microbiol.
28:1938-1941[Abstract/Free Full Text].
|
| 9.
|
Fenn, J. P.,
H. Segal,
B. Barland,
D. Denton,
J. Whisenant,
H. Chun,
K. Christofferson,
L. Hamilton, and K. Carroll.
1994.
Comparison of updated Vitek Yeast Biochemical Card and API 20C yeast identification systems.
J. Clin. Microbiol.
32:1184-1187[Abstract/Free Full Text].
|
| 10.
|
Gray, L. D., and G. D. Roberts.
1988.
Identification of medically important yeasts.
Clin. Microbiol. Newsl.
10:73-77.
|
| 11.
|
Hazen, K. C.
1995.
New and emerging yeast pathogens.
Clin. Microbiol. Rev.
8:462-478[Abstract].
|
| 12.
|
Kellogg, J. A.,
D. A. Bankert,
T. M. Brenneman,
M. A. Grove,
S. L. Wetzel, and K. S. Young.
1996.
Identification of clinical isolates of non-Enterobacteriaceae gram-negative rods using computer-assisted gas-liquid chromatography.
J. Clin. Microbiol.
34:1003-1006[Medline].
|
| 13.
|
Kellogg, J. A.,
D. A. Bankert, and V. Chaturvedi.
1998.
Limitations of the current Microbial Identification System for identification of clinical yeast isolates.
J. Clin. Microbiol.
36:1197-1200[Abstract/Free Full Text].
|
| 13a.
| Kellogg, J. A., D. A. Bankert, and V. Chaturvedi. 1999. Unpublished data.
|
| 14.
|
Merz, W. G.,
J. E. Karp,
D. Schron, and R. Saral.
1986.
Increased incidence of fungemia caused by Candida krusei.
J. Clin. Microbiol.
24:581-584[Abstract/Free Full Text].
|
| 15.
|
Meyer, R. D.
1992.
Current role of therapy with amphotericin B.
Clin. Infect. Dis.
14(Suppl.):154-160.
|
| 16.
|
Osterhout, G. J.,
V. H. Shull, and J. D. Dick.
1991.
Identification of clinical isolates of gram-negative nonfermentative bacteria by an automated cellular fatty acid identification system.
J. Clin. Microbiol.
29:1822-1830[Abstract/Free Full Text].
|
| 17.
| Paisley, R. (MIDI, Inc.) 1998. Personal
communication.
|
| 18.
|
Pfaller, M. A.,
T. Preston,
M. Bale,
F. P. Koontz, and B. A. Body.
1988.
Comparison of the Quantum II, API Yeast Ident, and AutoMicrobic systems for identification of clinical yeast isolates.
J. Clin. Microbiol.
26:2054-2058[Abstract/Free Full Text].
|
| 19.
|
Pinjon, E.,
D. Sullivan,
I. Salkin,
D. Shanley, and D. Coleman.
1998.
Simple, inexpensive, reliable method for differentiation of Candida dubliniensis from Candida albicans.
J. Clin. Microbiol.
36:2093-2095[Abstract/Free Full Text].
|
| 20.
|
Ramani, R.,
S. Gromadzki,
D. H. Pincus,
I. F. Salkin, and V. Chaturvedi.
1998.
Efficacy of API 20C and ID 32C systems for identification of common and rare clinical yeast isolates.
J. Clin. Microbiol.
36:3396-3398[Abstract/Free Full Text].
|
| 21.
|
Remel
1993.
Technical information. Sabouraud dextrose agar pH 5.6. TI no. 1768-A.
Remel, Lenexa, Kans.
|
| 22.
|
Sasser, M.
1991.
MIS whole cell fatty acid analysis by gas chromatography.
Microbial ID, Inc., Newark, Del.
|
| 23.
|
Warren, N. G., and K. C. Hazen.
1995.
Candida, Cryptococcus, and other yeasts of medical importance, p. 723-737.
In
P. R. Murray, E. J. Baron, M. A. Pfaller, F. C. Tenover, and R. H. Yolken (ed.), Manual of clinical microbiology, 6th ed. American Society for Microbiology, Washington, D.C.
|
| 24.
|
Wassertheil-Smoller, S.
1990.
Biostatistics and epidemiology. A primer for health professionals, p. 40-41.
Springer-Verlag, New York, N.Y.
|
| 25.
|
Wingard, J. R.
1995.
Importance of Candida species other than C. albicans as pathogens in oncology patients.
Clin. Infect. Dis.
20:115-125[Medline].
|
Journal of Clinical Microbiology, June 1999, p. 2080-2083, Vol. 37, No. 6
0095-1137/99/$04.00+0
Copyright © 1999, American Society for Microbiology. All rights reserved.
This article has been cited by other articles:
-
Peltroche-Llacsahuanga, H., Schmidt, S., Seibold, M., Lütticken, R., Haase, G.
(2000). Differentiation between Candida dubliniensis and Candida albicans by Fatty Acid Methyl Ester Analysis Using Gas-Liquid Chromatography. J. Clin. Microbiol.
38: 3696-3704
[Abstract]
[Full Text]
Top
Abstract
Text
