Candida albicans masquerading as gram-negative bacilli in the clinical laboratory.

Abstract

We report misidentification of Candida albicans as Gram-negative bacilli owing to colony morphology on MacConkey agar and subsequent inoculation into GN-ID/VITEK-2. ATCC and clinical Candida strains (n = 24) masqueraded as various bacterial species when experimentally inoculated into GN-ID cards. This phenomenon should be considered when peculiar taxa or susceptibility are encountered.

Full text

Page 1

SHORT COMMUNICATION
Candida albicans masquerading as Gram-negative bacilli in the
clinical laboratory
JACOB GILAD1, MICHAEL GILADI1,2& DAVID SCHWARTZ1
From the1Clinical Microbiology Laboratory and2Infectious Disease Unit, Tel-Aviv Sourasky Medical Centre, Tel-Aviv, Israel
Abstract
We report misidentification of Candida albicans as Gram-negative bacilli owing to colony morphology on MacConkey agar
and subsequent inoculation into GN-ID/VITEK-2. ATCC and clinical Candida strains (n?24) masqueraded as various
bacterial species when experimentally inoculated into GN-ID cards. This phenomenon should be considered when peculiar
taxa or susceptibility are encountered.
Introduction
Automated systems have revolutionized microorgan-
ism identification (ID) work schemes in clinical
microbiology practiceandavailablecommercialauto-
mated systems demonstrate high rates of accuracy in
ID [1]. Occasionally, microorganisms may be mis-
identified by automated systems because of technical
failure, limited number of available biochemical
reactions or imperfections in computerized ID algo-
rithms. Misidentification has been known to occur
since the introduction of automated ID systems [2]
and still occurs despite continuous product improve-
ment [1].
Whereas misidentification usually involves related
taxa, to our knowledge there are no reports of
misidentification of yeasts as bacteria with auto-
mated systems. Here we explore the misidentifica-
tion of Candida albicans as Gram-negative bacilli
(GNB) when inoculated into VITEK-2 (bioMer-
ieux, Marcy l’Etoile, France) during laboratory
work-up of clinical samples as well as experimental
inoculation of various yeast strains grown on differ-
ent media and discuss various aspects of this mishap.
Material and methods
Clinical specimen
A swab specimen was sent to the laboratory in
transport medium, inoculated on blood agar (BAP)
and MacConkey (MAC) plates (Hy-Labs, Rehovot,
Israel) and incubated at 358C in 5% CO2. Pre-
sumptive identification of GNB was performed by
inoculating the VITEK-2 (version 4.02) GN-ID
(colorimetric GNB identification) and AST-GN09
(standard GNB antimicrobial susceptibility panel)
cards according to manufacturer’s instructions.
Yeasts were identified using standard methods and
by inoculating the VITEK-2 YST-ID (yeast identi-
fication) card. Purity was ensured by inoculation of
quality control plates.
Experimental protocol
In order to elaborate the erroneous identification of
Candida species as GNB, 4 reference strains were
studied, including C.
C. krusei ATCC 6258, C. tropicalis ATCC 750
and C. parapsilosis ATCC 22019. Strains were
inoculated on Saboraud-Dextrose agar (SDA, Hy-
Labs, Rehovot, Israel) and incubated for 48 h. To
simulate the clinical case described below, growing
isolates were passed onto BAP (Hy-Labs, Rehovot,
Israel) as well as MAC produced by 2 different
manufacturers (Novamed, Jerusalem, Israel; Hy-
Labs, Rehovot, Israel), and incubated for 48 h at
358C in 5% CO2. The above experiments were
performed in triplicate.
Additionally, 20 random yeast isolates recovered
from hospitalized patients with fungaemia during
albicansATCC90023,
Correspondence: J. Gilad, Microbiology Laboratory, Tel-Aviv Sourasky Medical Centre, 6 Weizman St., Tel-Aviv 64239, Israel. Tel: ?972 578125093.
E-mail: giladko@014.net.il
Scandinavian Journal of Infectious Diseases, 2007; 39: 907?910
(Received 13 March 2007; accepted 17 April 2007)
ISSN 0036-5548 print/ISSN 1651-1980 online # 2007 Taylor & Francis
DOI: 10.1080/00365540701402996

Page 2

2006 corresponding to the 4 Candida species
analysed (5 strains of each species) were studied.
Isolates were identified based on standard mycolo-
gical procedures and inoculation of the VITEK-2
YST-ID card according to the manufacturer’s re-
commendations, and stored at ?708C until tested.
Clinical strains were similarly inoculated on SDA,
BAP and both MAC plates and further incubated.
Growth of yeast colonies on MAC was enumer-
ated as previously suggested according to the num-
ber of agar plate quadrants in which growth was
observed [3]. Growing isolates were inoculated into
VITEK-2 GN-ID cards from SDA, BAP and MAC
(4 ATCC strains) or from MAC only (20 clinical
strains), each after 48 h of incubation according to
the manufacturer’s instructions for GNB. Results of
the ID process were recorded, including positive and
negative biochemical reactions as well as the final ID
and its probability.
Results
Case report
The patient was a 2-y-old girl who had undergone
right leg amputation following a crush injury.
Exudate from a wound infection of the amputation
stump was swabbed by the attending physician and
sent to the clinical microbiology laboratory for
culture. Direct microscopy was not requested. Fol-
lowing incubation, pure growth of small round
pinkish colonies was evident on MAC. The isolate
was identified by VITEK-2 as Aeromonas sobria
(85.1% probability ? ‘acceptable identification’).
Antimicrobial susceptibility testing revealed non-
susceptibility to all tested agents. Examination of
quality control plates revealed pure growth and the
isolate was thus re-tested. This time, the isolate was
identified as Sphingomonas paucimobilis (89.6%
probability ? ‘good identification’), but was again
non-susceptible to all antimicrobials tested. At this
point, Gram stain and microscopy were performed,
revealing abundant yeasts but no bacteria. The
yeasts were identified as C. albicans.
Experimental findings
All 4 reference Candida strains grew well on SDA
and BAP, and ID was attempted for growing
colonies from all 24 plates (Table I). A single GNB
taxon was presumptively identified on 12 occasions
(50%), 2 to 3 taxons (‘low discrimination’) on 7
occasions (29%), and no identification on 5 occa-
sions (21%), the latter involving mostly C. tropicalis.
Notably, C. albicans was repeatedly identified as
S. paucimobilis (4 of 6 experiments) and C. krusei
Table I. Presumptive identification of control Candida strains analysed from Saboraud-dextrose and sheep blood agar in triplicate using the GN-ID VITEK-2 carda.
Saboraud-dextrose agar
Sheep blood agar
Strain
1
2
3
1
2
3
Candida. Albicans
(ATCC 90023)
Sphingomonas paucimobilis
(89.6%)
S. paucimobilis
(89.6%)
B. cepacia (93.2%)
S. paucimobilis (93.9%)
S. paucimobilis (93.3%)
Unidentified
Candida Krusei
(ATCC 6258)
Acinetobacter haemolyticus
(33%)/
Pseudomonas aeruginosa
(33%)/
Pseudomonas fluorescens
(33%)
P. fluorescens (99%)
P. fluorescens (99%)
A. haemolyticus (33%)/
P. aeruginosa (33%)/ P. fluorescens (33%)
A. haemolyticus (33%)/
P. aeruginosa (33%)/ P. fluorescens (33%)
A. haemolyticus (33%)/
P. aeruginosa (33%)/P. fluorescens (33%)
Candida tropicalis
(ATCC 750)
Burkholderia cepacia (94.9%) B. cepacia (94.9%)
Unidentified
Unidentified
Unidentified
Unidentified
Candida parapsilosis
(ATCC 22019)
A. haemolyticus (33%)/
P. multocida (33%)/P. fluorescens (33%)
Pasteurella multocida
(90.2%)
P. fluorescens (90.2%)
Pseudomonas stutzeri
(98.5%)
Aeromonas salmonicida
(50%)/
P. fluorescens (50%)
A. salmonicida (33%)/
P. fluorescens (33%)/
CDC group EF-4 (33%)
aProbabilities for identification calculated by the VITEK-2 software (version 4.02) appear in parentheses.
908
J. Gilad et al.

Page 3

was identified as Pseudomonas fluorescens in 2
experiments and yielded a low discrimination result
consistent with P. fluorescens/Acinetobacter haemo-
lyticus/Pseudomonas aeruginosa in 4 experiments.
With true GNB, the latter phenotype is resolved
according to manufacturer’s instructions by referring
to cytochrome oxidase activity and pyocyanin pro-
duction. C. parapsilosis yielded a greater diversity of
possible ID yielding various ID combinations invol-
ving 6 different taxa.
Both control and clinical strains showed variable
growth on MAC (Table II). Of 48 MAC plates
inoculated, Candida strains failed to grow only on 9
(18.7%) plates, 6 of which were C. tropicalis. While
C. krusei and C. tropicalis resulted in lack of
identification on most occasions, C. parapsilosis
strains yielded a diverse ID profile and C. albicans
universally yielded an identification of S. paucimo-
bilis from MAC (Table II). Misidentification of C.
albicans as S. paucimobilis occurred with varying
probabilities owing to a slight variability in biochem-
ical reactions. Of 47 reactions in the GN-ID card, 13
were positive with at least 1 C. albicans strain. All
strains were positive for 5 reactions (glucose, mal-
tose, mannose, sucrose and courmarate) and over
50% were positive for mannitol and citrate. Only
occasional strains were positive for L-proline aryla-
midase, beta-N-acetyl-galactosaminidase or alpha
glucosidase.
Discussion
C. albicans is a common pathogen that is readily
isolated and identified in clinical samples. On most
occasions, C. albicans grows on BAP within 24?48 h
and displays typical colonial morphology, at which
point yeasts are usually suspected and subsequently
confirmed by Gram staining. The presented case
and experimental results thus have several features
that warrant further discussion.
MAC is not intended to support the growth of C.
albicans and is erroneously regarded by many as
capable of inhibiting fungal growth although yeasts
may occasionally grow on MAC [4]. We were able to
grow Candida species on MAC plates from different
manufacturers in most experiments, although differ-
ences in rate and density of growth were observed,
suggesting variation in medium composition that
affects fungal growth. Interestingly, current recom-
mendations for quality control of MAC do not
mandate testing for fungal inhibition [5] and fungal
inhibition is usually not included in the manufac-
turer’s quality control reports provided with new lots
or batches. This limitation of MAC should thus be
taken into consideration in daily practice.
Ideally, Gram stain should be performed with
every specimen demonstrating growth. However,
this cannot be performed in the ‘real world’ due to
resource constraints. Current guidelines suggest that
in the routine workflow, Gram stain and microscopy
are not obligatory when growth typical of GNB is
seen on MAC [3]. Performance of Gram stain in our
case would have revealed the identity of microorgan-
isms at an earlier stage, but was not performed
according to the technologists’ judgment and la-
boratory workflow. However, Gram stain should
have been carried out after the first ID attempt
that yielded an inconceivable result.
Misidentification is not rare with all types of
automated ID systems and may involve Gram-
positive, Gram-negative or yeast isolates. With
VITEK-2, misidentification has been reported in
both the fluorometric and the newer colorimetric
methods. Anecdotal reports include misidentifica-
tion of various Aeromonas species as Vibrio species
[6,7], of Enterococcus faecium as Enterococcus
gallinarum or casseliflavus [8], of Enterbacter aero-
genes as Klebsiella pneumoniae [9], or of coagulase-
negative staphylococci as Kocuria spp. [10]. Overall,
the misidentification rate with colorimetric cards has
Table II. Presumptive identification of 24 Candida strains grown on MacConkey plates using the GN-ID VITEK-2 carda.
Growth (no. of isolates)b
Species NoneScant to ?1
?2 to ?4 Presumptive identification (no. of isolates)
Candida albicans
Candida krusei
1
0
8
0
3S. paucimobilis (n?6)
A. haemolyticus/P. aeruginosa/P. fluorescens (n?1);
unidentified (n?5)
S. paucimobilis (n?1); B. cepacia (n?1);
no growth (n?1); unidentified (n?3)
E. coli (n?3); P. fluorescens (n?1);
E. coli/P. multocida (n?1); no growth (n?1)
12
Candida tropicalis660
Candida parapsilosis228
aFor each species, one ATCC strain and 5 clinical strains were analyzed on MAC plates from two different manufacturers (total 48
experiments).
bEnumeration of growing colonies on agar was done as previously recommended [3].
Candida masquerading as Gram-negative bacilli
909

Page 4

recently been reported as B1% for Gram-positive
isolates [11] or GNB [12]. Moreover, misidentifica-
tion of yeasts as other yeast species has been
reported as well with a frequency of 1.7?4.9%
[13?15].
Yeasts are not expected to generate ID of bacterial
genera simply because fungi and bacteria are ana-
lysed with different cards. It is hoped that inad-
vertent inoculation of a microorganism to an
incompatible card would yield an inconclusive result
that will prompt re-analysis, but such situation is not
considered in the VITEK-2 manual, or in the
literature. Apparently, C. albicans as well as several
other Candida species react with ID cards designed
for GNB. Several biochemical reactions were uni-
formly positive with C. albicans while others yielded
variable results. This inconsistency with several
substrates may be related to the inoculation of cards
from different culture media, as it has been shown
that the source media influence ID results in
VITEK-2 [16]. BAP and SDA have also yielded
different results in our study. We speculate that this
may result from either an interference of medium
constituents with key biochemical reactions, or
medium influence on fungal growth and metabo-
lism, resulting in over- or under-expression of
various enzymes.
In conclusion, infectious disease specialists and
clinical microbiologists should be aware that inad-
vertentinoculationofanincompatibleVITEK-2card
may give rise to microorganism misidentification. We
speculate that a similar situation may occur in the
context of other microorganism-automated system
combinations and this should be investigated further.
‘Breakthrough’ growth of yeasts on MAC occurs with
commercial media, and could result in masquerading
of yeasts as GNB (especially if lactose is degraded).
However, this mishap should be suspected when a
peculiar organism is identified by a bacterial ID
system along with non-characteristic pan-resistance,
and prompt the performance of Gram stain and
microscopy.
References
[1] O’Hara CM. Manual and automated instrumentation for
identification of Enterobacteriaceae and other aerobic Gram-
negative bacilli. Clin Microbiol Rev 2005;
/18: /147?62.
[2] Stager CE, Davis JR. Automated systems for identification of
microorganisms. Clin Microbiol Rev 1992;
[3] York MK. Interpretation and rapid identification of bacterial
growth on primary culture media. In: Isenberg HD, ed.
Clinical Microbiology Procedures Handbook. 2nd edn.
Washington DC: ASM Press; 2004. 3.3.2.4.
[4] Winn WC JR., Allen S, Janda W, Koneman E, Procop G,
Schreckenberger P, et al. Koneman’ns Color Atlas and
Textbook of Diagnostic Microbiology. 6th ed. Baltimore,
MD: Lippincott Williams & Wilkins; 2006. p. 27.
[5] Jenkins SG, Sewell DL. Quality control. In: Isenberg HD,
ed. Clinical Microbiology Procedures Handbook. 2nd edn.
Washington DC: ASM Press; 2004. 14.2.10.
[6] Abbott SL, Seli LS, Catino M Jr, Hartley MA, Janda JM.
Misidentification of unusual Aeromonas species as members
of the genus Vibrio: a continuing problem. J Clin Microbiol
1998;
[7] Park TS, Oh SH, Lee EY, Lee TK, Park KH, Figueras MJ
et al. Misidentification of Aeromonas veronii biovar sobria as
Vibrio alginolyticus by the VITEK system. Lett Appl
Microbiol 2003;
[8] Garcia-Garrote F, Cercenado E, Bouza E. Evaluation of a
new system, VITEK 2, for identification and antimicrobial
susceptibility testing of enterococci. J Clin Microbiol 2000;
38: /2108?11.
[9] Claeys G, De Baere T, Wauters G, Vandecandelaere P,
Verschraegen G, Muylaert A, et al. Extended-spectrum
b-lactamase (ESBL) producing Enterobacter aerogenes phe-
notypically misidentified as
K. terrigena. BMC Microbiol 2004;
[10] Ben-Ami R, Navon-Venezia S, Schwartz D, Schlezinger Y,
Mekuzas Y, Carmeli Y. Erroneous reporting of coagulase-
negative staphylococci as Kocuria spp. by the VITEK 2
System. J Clin Microbiol 2005;
[11] Funke G, Funke-Kissling P. Performance of the new VITEK
2 GP card for identification of medically relevant Gram-
positive cocci in a routine clinical laboratory. J Clin Micro-
biol 2005;
[12] Rantakokko-Jalava K, Elo-Lehtonen E, Meurman O. Com-
parison of workflow and accuracy of identification and
antimicrobial susceptibility testing of clinical isolates of
Enterobacteriaceae, Pseudomonas aeruginosa and entero-
cocci by Vitek 2 and routine methods. APMIS 2006;
43?9.
[13] Graf B, Adam T, Zill E, Gobel UB. Evaluation of the
VITEK 2 system for rapid identification of yeasts and yeast-
like organisms. J Clin Microbiol 2000;
[14] Massonet C, van Eldere J, Vaneechoutte M, De Baere T,
Verhaegen J, Lagrou K. Comparison of VITEK 2 with
ITS2-fragment length polymorphism analysis for identifica-
tion of yeast species. J Clin Microbiol 2004;
[15] Meurman O, Koskensalo A, Rantakokko-Jalava K. Evalua-
tion of VITEK 2 for identification of yeasts in the clinical
laboratory. Clin Microbiol Infect 2006;
[16] Lowe P, Haswell H, Lewis K. Use of various common
isolation media to evaluate the new VITEK 2 colorimetric
GN card for identification of Burkholderia pseudomallei.
J Clin Microbiol 2006;
/5:
/302?27.
/36:
/1103?4.
/37: /349?53.
/
Klebsiella pneumoniaeor
/4:
/49.
/43:
/1448?50.
/43:
/84?8.
/114:
/
/38: /1782?5.
/42:
/2209?11.
/12:
/591?3.
/44:
/854?6.
910
J. Gilad et al.