Candida dubliniensis at a University Hospital in Saudi Arabia

Abstract

Candida dubliniensis is a newly described yeast species that is a close phylogenetic relative of C. albicans. Although it has been reported from different parts of the world, no detailed investigation of this species has been done in Saudi Arabia. The purpose of the present study was to identify C. dubliniensis isolates recovered from clinical specimens at a tertiary-care hospital in Riyadh, Saudi Arabia, and to determine the drug susceptibility profiles of those isolates. Over a period of 8 months, 823 germ tube- and chlamydospore-positive yeasts identified as C. albicans and recovered from different clinical specimens were screened for their ability to grow at 45°C on Sabouraud dextrose agar. Isolates which failed to grow at 45°C were presumptively identified as C. dubliniensis. The species identities were further confirmed by the production of pseudohyphae and chlamydospores on Staib agar and their inability to assimilate d-xylose and α-methyl-d-glucoside by using the API 20C AUX system. A total of 27 (3.3%) isolates were identified as C. dubliniensis. They were all recovered from 23 human immunodeficiency virus-negative patients. The prevalence of C. dublinensis in bronchoalveolar lavage (33.3%), oral (16.7%), and blood (16.7%) specimens was high. In addition, 33 isolates previously identified as C. albicans and preserved among our stock blood culture isolates were also recruited for the study. Of these, 5 isolates were found to be C. dubliniensis, thus making the total number of isolates identified as this species 32. Antifungal susceptibility testing of the C. dubliniensis isolates showed 100% sensitivity to amphotericin B, 97% sensitivity to each of fluconazole and ketoconazole, and 87.5% sensitivity to itraconazole. However, in contrast to other studies, the majority of the isolates (65.6%) showed high levels of resistance to flucytosine (MIC > 64 μg/ml). Further studies are warranted to investigate the cause of this unusually high rate of resistance to flucytosine of the C. dubliniensis isolates in this region.

Candida dubliniensis was first described as a novel species in 1995 by Sullivan et al. (31) and has been reported from various geographic locations (29, 30). Although previously associated with oral candidiasis in human immunodeficiency virus (HIV)-infected patients (5, 30, 31), it has more recently been recognized as a cause of superficial and systemic diseases in HIV-negative and HIV-positive patients (3, 4, 6, 10, 15, 23, 27). The close resemblance of C. albicans and C. dubliniensis has hampered the accurate and rapid identification of the latter from clinical specimens. In a recent retrospective study of the isolates in laboratory stock collections (isolates recovered between 1973 and 1994), approximately 2% of the isolates originally identified as C. albicans were found to be C. dubliniensis (19). However, the availability of numerous reliable and rapid tests, such as an inability to grow at 45°C (9, 11, 12, 22, 23), characteristic morphology on cornmeal agar (6, 11, 31), production of abundant pseudohyphae and chlamydospores on Staib agar (2, 28), formation of dark green colonies on CHROMagar Candida medium (5, 6, 29, 30), and carbohydrate assimilation tests (9, 21, 23, 29, 30) with the API 20C AUX and ID 32C systems (bioMerieux, Marcy l'Etoile, France), have greatly facilitated the differentiation of C. dubliniensis and C. albicans. In addition, the use of molecular techniques has significantly revolutionized investigations of various aspects of this yeast pathogen (8).

Although C. dubliniensis and C. albicans isolates are both susceptible to azoles, fluconazole resistance has been observed in clinical isolates of C. dubliniensis from AIDS patients with prior exposure to fluconazole (16, 17, 20). In addition, it has been shown that isolates of C. dubliniensis, unlike those of C. albicans, can develop stable resistance to fluconazole upon exposure in vitro (16, 17), suggesting a need to investigate the sensitivities of isolates of this species from various regions.

The present study was undertaken to determine the prevalence of C. dubliniensis isolates in different clinical specimens recovered at a tertiary-care hospital in Riyadh, Saudi Arabia. In addition, testing of the in vitro susceptibilities to different antifungal agents was done for all isolates identified as C. dubliniensis.

MATERIALS AND METHODS

During an 8-month period (between September 2000 and April 2001), we obtained 823 isolates of C. albicans from various clinical specimens of patients attending King Khalid University Hospital, Riyadh, Saudi Arabia. They were identified as C. albicans on the basis of a positive germ tube test result, production of chlamydospores in cornmeal agar, and resistance to cycloheximide when they were grown on cycloheximide (500 μg/ml)-containing medium (1). These isolates were reexamined to explore if any of them were actually C. dubliniensis and not C. albicans. In addition, 33 C. albicans isolates from our stock cultures from blood collected from January 1996 to December 2000 were also included in the study, making the total number of isolates examined 856.

Identification as C. dubliniensis.

All 856 isolates previously identified as C. albicans were screened for their ability to grow on Sabouraud dextrose agar when they were incubated at 45°C for 48 h. Yeasts that did not grow at 45°C were regarded as possible C. dubliniensis isolates (12, 22). They were then retested for germ tube production in human serum after 3 h of incubation at 37°C and were also cultured on cornmeal agar supplemented with 1% (wt/vol) Tween 80 for detection of production of hyphae and chlamydospores. The isolates were further subcultured on CHROMagar Candida medium (CHROMagar, Paris, France) and incubated at 37°C for 48 h for detection of the development of colored colonies (6, 11, 19, 31). Two isolates of C. albicans (ATCC 10231 and ATCC 90028) were also included as controls in all experiments.

Confirmation of the identities of the presumptively identified C. dubliniensis isolates was done by culturing them on Staib agar (28) containing 50 g of Guizotia abyssinica seed (pulverized), 1 g of glucose, 1 g of KH₂PO₄, 1 g of creatinine, and 15 g of agar per liter and incubation at 30°C for 72 h. The appearance of the colonies on the plates was inspected daily both macroscopically and microscopically with a ×10 objective (28). Additional confirmation of these isolates as C. dubliniensis was done by substrate assimilation profile analysis with the API 20C AUX yeast identification system (bioMerieux), according to the instructions of the manufacturer, with an inoculum derived from 48-h Sabouraud dextrose agar cultures. The results were recorded along with their corresponding codes, and a complete profile number was established for each isolate.

Serotyping.

As previous studies have shown that C. dubliniensis isolates belong exclusively to C. albicans serotype A (5, 6, 31), all C. dubliniensis isolates were serotyped on the basis of their agglutination reactions with antisera raised against Candida antigenic factors 6 and 13b (Iatron Laboratories, Inc., Tokyo, Japan). Each isolate was freshly subcultured, as specified in the instructions of the manufacturer, and was tested by slide agglutination. C. albicans control strains ATCC 10231 (serotype A) and ATCC 90028 (serotype B) were also tested (31).

Antifungal susceptibility testing.

Isolates of C. dubliniensis were tested for their susceptibilities to amphotericin B (E. R. Squibb & Sons Ltd., Hounslow, England), flucytosine (5FC; Sigma), fluconazole (Pfizer Ltd., Sandwich, England), and ketoconazole and itraconazole (Janssen Pharmaceutica, Beerse, Belgium). Broth macrodilution testing was performed according to the guidelines of the National Committee for Clinical Laboratory Standards (NCCLS) (18) with RPMI 1640 medium buffered to pH 7.0 with 0.165 M morpholinopropanesulfonic acid (MOPS) buffer. The stock drug solutions used for determination of MICs were made in either sterile distilled water or dimethyl sulfoxide, and 0.1 ml was inoculated into each tube. The yeasts at final concentrations of 0.5 × 10³ to 2.5 × 10³ cells per ml (by use of McFarland turbidity standards) were incubated in air at 35°C for 48 h with twofold dilutions of amphotericin B (0.03 to 16 μg/ml), fluconazole (0.125 to 64 μg/ml), ketoconazole (0.03 to 16 μg/ml), itraconazole (0.03 to 16 μg/ml), or 5FC (0.125 to 64 μg/ml). The MIC breakpoints were recorded at 24 and 48 h and were interpreted according to the suggestions of an NCCLS subcommittee (18) specifically for C. albicans and those of Rex et al. (25). The MICs of amphotericin B and 5FC for the strains were defined as the lowest concentration of drug that completely (100%) inhibited growth with a score of 0, whereas the MICs of the azoles were the lowest concentrations of the drug that produced an 80% reduction in turbidity with a score of 1+ in comparison to the growth for the drug-free control.

In view of the unusually high rate of resistance to 5FC observed by the broth macrodilution method, further testing was done by the Etest (AB Biodisk, Solna, Sweden). The Etest is based on the use of a plastic strip impregnated with a concentration gradient of the appropriate drug. Strips containing 5FC ranging in concentration from 0.002 to 32 μg/ml were placed on the surfaces of agar plates containing RPMI 1640 medium, MOPS, and 2% glucose. A pure colony of the test organism was inoculated into sterile distilled water to achieve a turbidity equivalent to that of a 0.5 McFarland standard. The agar plates were inoculated by use of a nontoxic cotton swab. The moisture was allowed to dry for 10 min, and the Etest strips were placed onto the agar surface. The plates were incubated at 35°C for 24 h or until sufficient growth was obtained to determine the endpoint. The results were read and interpreted according to the guidelines of the manufacturer. The MIC was defined as the lowest concentration of antifungal agent (5FC) at which there was 100% inhibition of organism growth. Quality control was ensured for both methods by including quality control strain C. albicans ATCC 90028, as recommended by NCCLS.

The interpretive criteria (MICs) used to consider an isolate drug sensitive were as follows: for fluconazole, <64 μg/ml; for itraconazole, <1.0 μg/ml; for ketoconazole, <1.0 μg/ml; for amphotericin B, ≤1.0 μg/ml; and for 5FC, <8.0 μg/ml. These were based on studies with C. albicans published previously (18, 24, 25).

RESULTS

Among the clinical specimens, 32 of the 823 germ tube- and chlamydospore-positive yeast isolates which failed to grow at 45°C were presumptively identified as C. dubliniensis. On the bases of further phenotypic tests, 27 of the 32 isolates (84.4%) recovered from 23 HIV-negative patients were confirmed to be C. dubliniensis, whereas the remaining 5 isolates were confirmed to be C. albicans. Of the 33 blood culture stock isolates, 5 isolates recovered from HIV-negative patients were identified as C. dubliniensis.

The prevalence of C. dubliniensis in clinical specimens is shown in Table 1. The majority of isolates were recovered from bronchoalveolar lavage specimens (33.3%), followed by blood specimens (16.7%), oral specimens (16.7%), and other specimens (tracheal aspirate, stool, urine, skin, and high vaginal swabs) (Table 1). Details about the 32 C. dubliniensis isolates (27 isolates recovered from 23 patients and 5 isolates recovered from stock blood cultures) are presented in Table 2.

TABLE 1.

Frequency of C. dubliniensis isolates among yeasts identified as C. albicans recovered from various clinical specimens

Clinical specimen	No. (%) of yeasts
	Totala	C. albicans	C. dubliniensis
High vaginal swab	430	422 (98.1)	8 (1.9)
Urine	210	202 (96.2)	8 (3.8)
Skin	75	73 (97.3)	2 (2.7)
Sputum	40	39 (97.3)	1 (2.5)
Tracheal aspirate	24	23 (95.8)	1 (4.2)
Bronchoalveolar lavage	9	6 (66.7)	3 (33.3)
Oral	6	5 (83.3)	1 (16.7)
Throat swab	4	4 (100)	0
Blood	12	10 (83.3)	2 (16.7)
Foley catheter tip	6	6 (100)	0
Body fluid (ascitic fluid + bile)	5	5 (100)	0
Feces	2	1 (50)	1 (50)
Total	823	796 (96.7)	27 (3.3)

^aNot included here are the 33 yeasts obtained from preserved stock blood culture isolates, of which 5 were found to be C. dubliniensis (see text for detail).

TABLE 2.

Clinical data and characteristics for C. dubliniensis recovered from clinical specimens^a

Isolate group and serial no.	Age (yr)/sex	Clinical condition	Specimen	APT 20C AUX system assimilation profile code	MIC (μg/ml) at 48 h
					5FC	FLU	ITR	KET	AMB
Clinical isolates
1	31/F	CBD obstruction	Urine	2142114	>64	1.0	1.0	≤0.03	0.125
2, 3	16/Fb	Laprotomy	Abdominal wound	2152114	≤0.125	0.25	0.06	≤0.03	0.25
4	45/M	UTI	Urine	6152114	>64	0.25	0.125	0.06	0.25
5	62/M	Lung abscess	BAL	2172014	>64	0.25	≤0.03	0.06	0.25
6	24/F	UTI	Urine	6142034	>64	0.25	0.25	0.125	0.25
7	16/M	Submandibular infection	Oral	6142134	≤0.125	2.0	1.0	0.5	0.25
8	21/F	Vaginal discharge + pregnancy	HVS	2140134	≤0.125	0.125	0.25	≤0.03	0.25
9	29/F	Vaginal discharge + pregnancy	HVS	2140134	≤0.125	>64	>16	4.0	0.5
10	4/M	Urinary incontinence + diarrhea	Feces	2142014	>64	≤0.125	0.125	≤0.03	0.25
11	33/F	Vaginal discharge + pregnancy	HVS	2172134	>64	0.5	0.25	0.125	0.25
12, 13	31/Fb	Vaginal discharge	HVS	2172034	>64	0.5	0.25	0.25	0.25
14	37/F	Vaginal discharge	HVS	2142034	≤0.125	0.25	≤0.03	0.125	0.25
15	24/F	Vaginal discharge	HVS	6172134	>64	0.5	0.25	≤0.03	0.25
16	36/F	UTI	Urine	6172034	>64	1.0	1.0	0.125	0.25
17, 18	70/Mb	Hepatocellular carcinoma	Blood	6172134	>64	0.5	0.125	0.125	0.25
19	44/F	UTI	Urine	2152134	>64	0.25	0.25	≤0.03	0.125
20	60/F	Sleep apnea + respiratory infection	Sputum	2172114	≤0.125	4.0	0.25	0.125	0.25
21	36/F	Pregnancy	Urine	2152134	>64	0.5	0.125	≤0.03	0.125
22, 23	27/Fb	UTI	Urine	2142134	>64	0.25	0.25	0.125	0.25
24	1/M	Pneumonia	Tracheal aspirate	6152134	≤0.125	0.25	0.125	0.125	0.25
25	50/F	Pleural effusion	BAL	2142114	>64	1.0	0.5	≤0.03	0.5
26	26/F	Vaginal discharge	HVS	2152134	≤0.125	1.0	0.25	≤0.03	0.25
27	39/F	Pneumonia	BAL	2152034	>64	0.25	0.125	≤0.03	0.25
Stock isolates recovered from blood cultures
28	35/F	Hepatosplenomegaly + ARF	Blood	2172134	≤0.125	0.5	0.5	0.125	0.5
29	40/F	Chronic renal failure + SLE	Blood	6172134	>64	1.0	0.25	0.125	0.25
30	12/M	ALL, fever	Blood	2172134	>64	1.0	0.25	0.25	0.25
31	25/M	ALL, fever	Blood	2152134	>64	0.5	0.5	0.25	0.25
32	1/F	Disseminated TB	Blood	2152134	≤0.125	2.0	0.5	0.25	0.5

^aAbbreviations: F, female; M, male; CBD, chronic bile duct; UTI, urinary tract infection; ARF, acute renal failure; ALL, Acute lymphocytic leukemia; BAL = bronchoalveolar lavage; HVS, high vaginal swab; SLE, systemic lupus erythematosus; TB, tuberculosis; FLU, fluconazole; ITR, itraconazole; KET, ketoconazole; AMB, amphotericin B.

^bPatients from whom C. dubliniensis was recovered from the same specimen on two occasions.

All 32 isolates which failed to grow at 45°C formed short hyphae and chlamydospores in triplets and contiguous pairs considered typical for C. dubliniensis on cornmeal agar and similar to those reported for C. dubliniensis by Sullivan et al. (31). However, both C. albicans ATCC 10231 and ATCC 90028 showed single terminal chlamydospores regarded as typical of C. albicans. Characteristic dark green-blue colonies were formed by all 32 isolates on CHROMagar Candida medium, suggestive of C. dubliniensis (6, 29, 31), whereas C. albicans control isolates ATCC 10231 and ATCC 90028 formed lighter green colonies on this medium, typical of C. albicans. Further confirmation of these isolates as C. dubliniensis was based on their formation of rough colonies with abundant hyphae and chlamydospores on Staib agar (28). However, the control C. albicans isolates formed smooth colonies without hyphae and chlamydospores. None of the 32 isolates assimilated xylose or α-methyl-d-glucoside, giving the Analytab Products (API) system profiles corresponding to the identification of C. dubliniensis (Table 2). On the other hand, the C. albicans control strains assimilated the two sugars, thus giving the API system code corresponding to the identification of C. albicans. Positive agglutination reactions with antiserum against factor 6 but not with antiserum against factor 13b were observed for all isolates (n = 32), and they were therefore grouped as serotype A.

Five isolates which did not grow at 45°C and which were considered suspicious proved to be C. albicans on further phenotypic tests. These isolates and the control C. albicans isolates formed long hyphae and produced the single terminal chlamydospores regarded as typical of C. albicans on cornmeal agar, formed the light green colonies on CHROMagar Candida medium considered typical for C. albicans, and formed smooth colonies that were similar to those of C. albicans on Staib agar and that on microscopic examination revealed the absence of hyphae, pseudohyphae, and chlamydospores. Positive assimilation of d-xylose and α-methyl-d-glucoside was recorded by using the API 20C AUX system and gave the API system code typical of C. albicans, thus confirming their identification.

Antifungal susceptibility testing.

Testing of the in vitro susceptibilities of the 32 isolates of C. dubliniensis revealed that the MICs of amphotericin B were within the range of 0.125 to 0.5 μg/ml. Ninety-seven percent of the C. dubliniensis isolates were susceptible to each of fluconazole (MIC range, ≤0.125 to 4.0 μg/ml) and ketoconazole (MIC range, ≤0.03 to 0.5 μg/ml), 87.5% were susceptible to itraconazole (MIC range, 0.03 to 0.5 μg/ml), but only 34.4% were susceptible to 5FC (MIC ≤0.125 μg/ml) (Table 2). One isolate recovered from a high vaginal swab was resistant to three antifungals, fluconazole, ketoconazole, and itraconazole.

The MICs of 5FC obtained by the Etest indicated that 65.6% of the C. dubliniensis isolates were resistant to 5FC, as no growth inhibition around the strip was observed (MIC > 32 μg/ml); however, 11 of 32 (34.4%) isolates showed growth inhibition around the strip below drug concentrations of 0.125 and <0.094 μg/ml, indicating the susceptibilities of these isolates to 5FC.

DISCUSSION

This is the first study highlighting the recovery of C. dubliniensis from different clinical specimens from Saudi Arabia. An inability to grow at 45°C was used as an initial screening method for the presumptive identification of C. dubliniensis among isolates previously identified as C. albicans. This is in agreement with the findings of other investigators, who have confirmed that no C. dubliniensis isolate grows at 45°C and that this test can be used to screen large numbers of germ tube- and chlamydospore-positive yeast isolates (11, 12, 22, 23). Several investigators have reported that determination of the inability of C. dubliniensis to grow at 45°C is a reliable test that can be used to differentiate it from C. albicans (9, 11, 12, 22, 23). However, five isolates that did not grow at 45°C failed to give further positive tests for C. dubliniensis and were identified as C. albicans. This indicates that failure to grow at 45°C is not sufficient to confirm the identity of C. dubliniensis (9, 11, 13).

In this study, all 32 isolates which produced abundant chlamydospores on cornmeal agar formed dark green colonies on CHROMagar Candida medium and hence were presumptively identified as C. dubliniensis. These results are in agreement with those in reports of several other investigators who have highlighted the use of these tests for the presumptive identification of C. dubliniensis (5, 11, 12, 19, 30, 31). However, since some investigators question the value of these tests (13, 26, 29), we did other confirmatory tests to identify the species, namely, tests for determination of the characteristic morphology on Staib agar medium and sugar assimilation.

The use of Staib agar proved useful and was used in one of the confirmatory tests to differentiate C. dubliniensis and C. albicans. Our findings confirm the results of other investigators, indicating the usefulness of Staib agar as a medium for differentiating the two closely related yeast species (2, 28).

None of the 32 isolates identified as C. dubliniensis by the methods described above assimilated α-methyl-d-glucoside or xylose, the two key sugars used to differentiate C. dubliniensis from C. albicans, indicating the potential of the sugar assimilation test for the differentiation of the two yeast species (9, 12, 21, 23). It has been proposed that the identities of germ tube- and chlamydospore-positive isolates which fail to grow at 45°C can be confirmed by carbohydrate assimilation with the API yeast identification system (11, 12, 21, 23, 31).

All the C. dubliniensis isolates tested in this study proved to be serotype A. This finding is in accordance with those of other studies which have shown that C. dubliniensis belongs exclusively to C. albicans serotype A, strongly suggesting that our 32 isolates were C. dubliniensis (6, 23, 29-31).

In the present study the recovery of C. dubliniensis from different clinical specimens of HIV-negative patients is highlighted, and our results are in agreement with those presented in earlier reports that described the presence of this novel species among HIV-negative patients (12, 15, 19, 23). The prevalence of C. dubliniensis observed in this study (3.3%) is considerably higher than that reported by Meis et al. (15) among HIV-negative patients (0.8%). However, a higher prevalence (20%) of C. dubliniensis has been reported in the oral cavities of HIV-positive patients (12). In our study a higher prevalence of C. dubliniensis was observed in bronchoalveolar lavage (33.3%) and oral (16.7%) specimens, thereby supporting the earlier findings that C. dubliniensis is an opportunistic pathogen predominately associated with colonization and infection of the oral cavity and upper respiratory tract (6, 12, 15, 23, 27, 29).

The recovery of C. dubliniensis from cultures of blood from patients with a variety of clinical conditions indicates that the clinical spectrum of C. dubliniensis is not different from that of C. albicans. Moreover, our results also indicate the ability of C. dubliniensis to cause invasive infection in HIV-negative patients with different clinical conditions, as has been reported earlier (4, 10). To the authors' knowledge, the recovery of this yeast from cultures of blood from patients with disseminated tuberculosis and systemic lupus erythematosus has not been reported previously.

The amphotericin B MICs for all C. dubliniensis isolates tested in this study were <1 μg/ml, which is in agreement with those provided in previous reports (13, 20, 24, 27). Although there is considerable concern about fluconazole resistance among clinical isolates of C. dubliniensis recovered from HIV-infected patients previously treated with the drug (16, 17, 20), we found that a very low percentage of C. dubliniensis isolates were resistant to fluconazole and ketoconazole. These results are in agreement with those of Odds et al. (19) and Pfaller et al. (20); however, Quindos et al. (24) reported significantly higher rates of resistance to fluconazole (17%) and ketoconazole (24%) in the non-HIV-infected patients. Resistance to itraconazole was seen in 12.5% of our isolates, which is almost similar to the 13.8% resistance rate indicated by Quindos et al. (24), whereas low resistance rates (0 and 4.6%) have been reported by other investigators (13, 20).

A high percentage (65.6%) of our C. dublinineisis isolates were resistant to 5FC in vitro. This is contrary to the results of other investigators, who did not observe 5FC resistance in C. dubliniensis (19, 20, 24, 27), although McCullough et al. (14) reported that these isolates were less susceptible to 5FC. It is worth mentioning that clinically 5FC is used in combination with other antifungals for the treatment of fungal infections (7). This study constitutes the first report of the isolation of C. dubliniensis from clinical specimens recovered in Saudi Arabia.