Presumptive Identification of Candida kefyr on Levine Formulation of Eosin Methylene Blue Agar

Abstract

Three hundred thirty-one yeast and yeast-like isolates were cultivated on eosin methylene blue agar. While the sensitivity rate for Candida kefyr isolates producing a metallic green sheen was 81.8%, the high positive predictive value (100%) for yeast isolates with this phenotype belonging to C. kefyr suggests that these isolates can be presumptively identified as C. kefyr.

Although not isolated with the same relative frequency as yeast species like Candida albicans, Candida glabrata, Candida parapsilosis, and Candida tropicalis (15), Candida kefyr (formerly Candida pseudotropicalis) has been viewed as an emerging fungal pathogen (4, 5). Past medical literature has described cases of systemic disease that were elucidated postmortem (3, 13, 17, 20) and cases of cystitis (6) with C. kefyr as their etiology. More recent reports document C. kefyr as a confirmed agent of esophagitis (9) and fatal fungemia with disseminated disease (11) while suggesting a role for this organism in diseases of the reproductive tracts of infertile women (1).

Rapid and accurate yeast identification schemes have dated back to the 1960s with the development of the 2-h germ tube test for identification of C. albicans (10). Other methods not requiring the incubation period necessary for carbohydrate assimilation testing include those based on detection of pertinent enzymatic activities (14, 16). Colorimetric metabolic phenotypes demonstrated on CHROMagar Candida (CHROMagar Company, Paris, France) medium may yield presumptive identifications of C. albicans and C. tropicalis (12), while distinctive growth characteristics on eosin methylene blue (EMB) agar provide presumptive identifications of C. glabrata (2, 7; R. Jones, R. Master, and D. Powell, Abstr. 35th Intersci. Conf. Antimicrob. Agents Chemother., abstr. D10, 1995). This study shows further usefulness of EMB agar in the presumptive identification of C. kefyr.

Three hundred thirty-one clinical isolates were obtained from collections maintained by the University of Iowa College of Medicine and the University of Wisconsin Medical School. Concordant identification of the isolates were previously made by two independent sources using a variety of methods, including germ tube formation, carbohydrate assimilation testing using commercial methods, and distinctive colorimetric phenotype on CHROMagar Candida medium. This representative collection consisted of one genus of yeast-like filamentous fungi (Geotrichum spp.), one genus of yeast-like algae (Prototheca spp.), and 28 species of yeast (including six species comprising three yeast holomorphs). The number of C. albicans, C. glabrata, C. parapsilosis, and C. tropicalis isolates used in this investigation attempted to reflect their relative frequency of isolation in clinical disease (15).

Isolates were cryptically encoded and cultivated on potato dextrose agar plates (Remel, Lenexa, Kans.) for 48 to 72 h in 35°C ambient air. Isolated colonies were then subcultured onto two EMB agar plates (Remel) containing lactose as the sole carbohydrate source (Levine formulation [8]). One plate was incubated in 35°C ambient air, while the other was incubated at 35°C in 5% CO₂. Two independent parties (observers A and B) observed colonial growth at 48 and 96 h of incubation for evidence of a metallic green sheen, with a subset of randomly selected isolates (n = 285) also assessed for evidence of precipitate deposition. Isolates observed to exhibit a metallic green sheen were subcultured onto potato dextrose agar. Following 24 to 48 h of incubation in 35°C ambient air, the identities of these isolates were reconfirmed by use of the Yeast Biochemical Card (bioMérieux, Marcy l'Etoile, France). Differences in the rates of observation of metabolic activity (represented by the development of a dark central precipitate or the production of a metallic green sheen) were analyzed by the significance test of proportions (18). The alpha level was set at 0.05 before the experiments were started.

The observation of colonies producing a dark central precipitate in the agar, was not a strong predictor of organism identification for three reasons. First, analysis of 285 randomly selected isolates determined that incubation conditions had a significant impact on the deposition of a precipitate by an isolate. At the 48-h time point, both observers reported the presence of approximately twice as many isolates with precipitate deposition when incubation was done in a CO₂-enriched atmosphere than when incubation was done in ambient air (P < 0.001) (Table 1). This trend held true when isolates were observed after 96 h (P < 0.001). Exogenous atmospheric acidity induced by the carbon dioxide environment may provide an explanation for this phenomenon.

TABLE 1.

Differences in interpretation of precipitate deposition of 285 randomly selected yeast and yeast-like species on EMB agar (Levine formulation)

Observer	% of isolates with precipitate deposition after incubation for:
	48 h		96 h
	Ambient air	CO2	Ambient air	CO2
A	20.0	36.1	44.9	64.2
B	10.9	21.8	37.9	54.7

Second, determination of precipitate deposition as an indicator of lactose fermentation, by yeast species on EMB medium may not be as straightforward as determination of patterns exhibited by coliforms of the family Enterobacteriaceae (8). When data recorded by observer A are compared to those of observer B for the 285-isolate subset, significant differences occurred in the percentage of isolates producing a precipitate at 48 h either in the presence (P < 0.001) or absence (P = 0.003) of CO₂ enrichment (Table 1). Differences in the data reported by the observers still existed at the 96-h time point, although only those differences from assays involving incubation in enriched CO₂ were significant (P = 0.021). A variety of sources (5, 7, 19) have actually documented a paucity of yeast species that are capable of fermenting lactose. Therefore, the precipitate observed in many of the isolates may actually be an artifact of the colony texture itself or may represent a topographical growth apex of the colony.

Finally, of 27 groups of organisms studied, 19 contributed at least one isolate that was deemed by both observers to exhibit a dark precipitate (Fig. 1A) following incubation in 35°C ambient air. Two yeast species, Candida inconspicua and Candida rugosa, contributed at least one isolate that produced a dark precipitate within colonial growth, as determined by only one of the observers. Isolates of six yeast or yeast-like species (Candida sake, Cryptococcus albidus, Cryptococcus neoformans, Cryptococcus terreus, Prototheca spp., and Saccharomyces spp.) failed to produce a dark precipitate (Fig. 1B).

FIG.1.

Representative photographs of yeast species cultivated on the Levine formulation of EMB agar exhibiting the presence of a dark central precipitate (C. albicans) (A), the absence of a dark central precipitate (C. glabrata) (B), and a distinctive metallic green sheen (C. kefyr) (C).

In contrast, growth of yeast isolates on EMB agar that resulted in a distinctive metallic green sheen (Fig. 1C) appeared to be a strong diagnostic predictor. Of 331 yeast or yeast-like isolates, 18 (5.4%) exhibited this phenotype (Table 2). These isolates were identified as C. kefyr (one isolate was the Kluyveromyces marxianus teleomorph of C. kefyr), and the identifications were reconfirmed via a commercial system. Fifteen of these 18 isolates (83.3%) exhibited the phenotype at 48 h of incubation. Concordance between both observers was achieved for 96-h readings of isolates cultivated both in ambient air and in an enriched CO₂ environment. More isolates incubated in 35°C ambient air exhibited the characteristic sheen than did those incubated in a CO₂-enriched environment, although this difference was not statistically significant when all study isolates (P = 0.862) or all C. kefyr isolates (P = 0.708) were considered.

TABLE 2.

Distribution of 331 yeast and yeast-like species and frequency of isolates exhibiting a metallic green sheen phenotype

Organism	n	No. (%) of isolates exhibiting metallic green sheen after incubation for:
		48 h		96 h
		Ambient air	CO2	Ambient air	CO2
Candida albicans	50	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida dubliniensis	10	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida famata	10	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida glabrata	27	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida guilliermondii	11	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida inconspicua	5	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida kefyra	22	14 (63.6)b	15 (68.2)	18 (81.8)	17 (77.3)
Candida krusei	23	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida lambica	1	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida lipolyticac	9	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida lusitaniae	13	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida parapsilosis	22	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida pelliculosad	10	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida rugosa	5	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida sake	1	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida stellatoidea	1	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida tropicalis	23	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Candida zeylanoides	1	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Cryptococcus albidus	1	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Cryptococcus humicola	5	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Cryptococcus laurentii	7	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Cryptococcus neoformans	22	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Cryptococcus terreus	2	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Geotrichum spp.e	11	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Prototheca spp.f	4	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Rhodotorula spp.g	12	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Saccharomyces spp.h	11	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Trichosporon spp.i	13	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Total	331	14 (4.2)	15 (4.5)	18 (5.4)	17 (5.1)

^aIncludes one isolate of Kluyveromyces marxianus teleomorph.

^bA second observer reported 15 (68.2%).

^cIncludes one isolate of Yarrowia lipolytica teleomorph.

^dIncludes seven isolates of Pichia anomala teleomorph.

^eIncludes species identified as Geotrichum candidum and other Geotrichum spp.

^fIncludes species identified as Prototheca wickerhamii and other Prototheca spp.

^gIncludes species identified as Rhodotorula rubra and other Rhodotorula spp.

^hIncludes species identified as Saccharomyces cerevisiae and other Saccharomyces spp.

ⁱIncludes species identified as Trichosporon asahii, Trichosporon beigelii, and other Trichosporon spp.

While the positive predictive value for a green sheen colony being C. kefyr was 100% in this study, the sensitivity of the EMB agar screen for all isolates of C. kefyr tested was 81.8%. It was unclear why four C. kefyr strains did not exhibit this phenotype. Some C. kefyr strains have differential lactose-metabolizing capabilities (5, 19). However, subsequent analysis of the four strains on the confirmatory commercial system showed that the isolates did possess the ability to utilize lactose.

In conclusion, while production of a dark precipitate on EMB agar (Levine formulation) may not be predictive of specific yeast species, data from this study suggest that isolation of yeast populations that produce a metallic green sheen can yield a presumptive identification of C. kefyr. This finding can augment the practice of laboratories that already incorporate this medium into certain workups, particularly those that use the medium to assist in the presumptive identification of C. glabrata.