Risk of Fungemia Due to Rhodotorula and Antifungal Susceptibility Testing of Rhodotorula Isolates

Abstract

Rhodotorula infections occur among patients with immunosuppression and/or central venous catheters. Using standardized methods (NCCLS M27-A), we determined the antifungal susceptibilities of 10 Rhodotorula bloodstream infection isolates. Patient information was collected for clinical correlation. The MICs of amphotericin B and posaconazole were the lowest, and the MICs of triazoles and echinocandins were higher than those of other antifungal agents.

Rhodotorula species have emerged as human pathogens due to immunosuppression and foreign-body technology. Forty-three cases of Rhodotorula bloodstream infections (BSIs) were reported between 1960 and 2000 (16). Risk factors include central venous catheters (CVCs) and malignancies (5, 12, 14).

Lack of standardization for susceptibility testing and a paucity of cases hamper treatment recommendations. Results of the susceptibility testing of 66 Rhodotorula isolates are available, but most of the isolates were not tested according to NCCLS guidelines (13). Previous data show amphotericin B to have the most favorable MIC. For the newer antifungal agents, data have been reported only for voriconazole and posaconazole (total, 10 isolates) (2, 6, 7).

A standardized method for antifungal susceptibility testing of yeasts (Candida and Cryptococcus species) has been defined by the NCCLS (5). Rhodotorula species, like Cryptococcus species, are heterobasidiomycetes and thus might be reliably tested by this protocol. We performed tests on clinical Rhodotorula isolates to determine susceptibility to traditional antifungal agents and to voriconazole, posaconazole, caspofungin, and micafungin.

Materials and methods.

Manufacturers provided powder of amphotericin B (Bristol Myers-Squibb, New York, N.Y.), flucytosine (Hoffman-LaRoche, Nutley, N.J.), voriconazole and fluconazole (Pfizer, New York, N.Y.), itraconazole (Janssen Pharmaceuticals, Beerse, Belgium), posaconazole (Schering Plough, Kenilworth, N.J.), caspofungin (Merck, Whitehouse Station, N.J.), and micafungin (Fujisawa Pharmaceutical Co., Ltd., Osaka, Japan).

Patients with Rhodotorula BSI at Duke University Medical Center from 1992 to 2001 were identified. Frozen specimens were obtained from the laboratory specimen bank, identified by using standard microbiologic techniques (API 20 C Aux; Biomerieux, Marcy L'Etoile, France), and subcultured at least twice onto Sabouraud dextrose agar to ensure viability and purity.

Susceptibility testing utilized the NCCLS M27-A macrodilution method.

Stock suspensions were made from 48-h cultures of isolates on Sabouraud dextrose agar at 35°C (30°C for two isolates). Turbidities were spectrophotometrically adjusted, and the suspensions were diluted 1 to 100 and then 1 to 20 in RPMI medium, resulting in a concentration of 0.5 × 10³ to 2.5 × 10³ cells per ml.

Serial dilutions from standard drug stocks were prepared within 24 h of testing and stored at 4°C until use.

Each tube containing 0.1 ml of drug solution was inoculated with 0.9 ml of inoculum suspension. Candida parapsilosis ATCC 22019 was tested as a control with each set. The tubes were incubated without agitation at 30° and 35°C until growth was sufficient (72 h) for the determination of the MIC. The MIC resulting in 100% growth inhibition (MIC₁₀₀) was determined for amphotericin B. For the other antifungal agents, the MIC₁₀₀ and MIC₈₀ were determined.

Patient data were abstracted from computerized records and charts. The study was approved by the Institutional Review Board at Duke University Medical Center.

MIC ranges and the MIC₉₀ were obtained. Descriptive statistics were used to evaluate patient data.

Results.

Of the 10 Rhodotorula isolates identified, 8 were Rhodotorula mucilaginosa (synonyms, R. rubra and R. pilimanae). Two were Rhodotorula glutinis. Except for isolates 3 and 4, all produced sufficient growth at 35°C to determine the MICs of the antifungal agents at 72 h. The MICs for isolates 3 and 4 were determined on day 5 of growth at 30°C.

The MICs of all drugs tested for the control isolate are listed in Table 1 and were comparable to the expected ranges (3).

TABLE 1.

Susceptibilities of clinical isolates of Rhodotorula species to antifungal agents

Yeast (no. of strains)	Antimicrobial agent	MIC rangea (μg/ml)
R. mucilaginosa (8)	Amphotericin B	0.25-1.0
	Flucytosine	0.125-0.25
	Fluconazole	32->64
	Itraconazole	0.5-4.0
	Voriconazole	1->8.0
	Posaconazole	0.5-2.0
	Caspofungin	16->16
	Micafungin	>64
R. glutinis (2)	Amphotericin B	0.5-1.0
	Flucytosine	0.125-0.25
	Fluconazole	32->64
	Itraconazole	1.0-4.0
	Voriconazole	4.0-8.0
	Posaconazole	1.0-2.0
	Caspofungin	16->16
	Micafungin	>64

^aAmphotericin is reported as MIC₁₀₀. All other values are reported as MIC₈₀.

The susceptibilities of Rhodotorula isolates at 35°C are shown in Table 1 (strains 3 and 4 at 30°C). The isolates were most susceptible to amphotericin B (MIC₉₀, 1.0 μg/ml) and flucytosine (MIC₉₀, 0.25 μg/ml) and less susceptible to azoles. The MICs of caspofungin (MIC₉₀, 16 μg/ml) and micafungin (MIC₉₀, >64 μg/ml) were high.

Patient data and treatment regimens are shown in Table 2. To the best of our knowledge, no patient had a recurrence after therapy or died as a direct result of Rhodotorula fungemia.

TABLE 2.

Clinical characteristics of patients with Rhodotorula fungemia^a

Patient	Age	Speciesb	No. of positive cultures	Underlying disease	Cause of immunosuppression	Transplant	Antifungal agent(s) used	Line removed
1	23	R. mucilaginosa	4	HIV-intravenous drug use	CD4 < 50	None	Fluconazole then amphotericin Bc	No
2	7	R. glutinis	2	Glioblastoma multiforme, TPNd	Methylprednisolone	Auto BMTd	Amphotericin B	No
3	43	R. mucilaginosa	1	Acute leukemia	Neutropenia	None	None	Yes
4	1 month	R. mucilaginosa	2	Congenital heart disease, TPN	None	None	Amphotericin B	Unknown
5	54	R. mucilaginosa	1	TPN, Crohn's disease	Tacrolimus, prednisone, azathioprine	Kidney	Fluconazole	Yes
6	29	R. glutinis	1	TPN, short gut	None	None	Fluconazole	Yes
7	4	R. mucilaginosa	1	Neuroblastoma	Neutropenia	None	Amphotericin B, 1 dose	No
8	32	R. mucilaginosa	2	Gastroparesis, TPN	None	None	Fluconazole	Yes
9	31	R. mucilaginosa	1	Lung transplant, cystic fibrosis	Prednisone, tacrolimus	Lung	ABLCd	Yes
10	35	R. mucilaginosa	3	Sickle cell disease	None	None	Fluconazole then amphotericin B	Yes

^aAll patients had a central line at the time of fungemia.

^bRhodotorula rubra and pilimanae were changed to Rhodotorula mucilaginosa.

^cThe patient received one dose of fluconazole (400 mg), then 2 doses of amphotericin B (0.7 mg/kg of body weight) and then left the hospital against medical advice.

^dTPN, total parenteral nutrition; BMT, bone marrow transplant; ABLC, amphotericin B lipid complex.

Discussion.

Similar to previous findings with NCCLS broth macrodilution (2, 7) and Sensititre YeastOne testing (8), the isolates appeared more susceptible to amphotericin B and flucytosine than to azoles and echinocandins. The MICs of posaconazole were better than previously reported (range, 2.0 to >4.0 μg/ml for 7 isolates) (2). The MIC range of voriconazole reported for 5 isolates tested by the microdilution method was 0.25 to 4 μg/ml, lower than our MIC range of 1 to >8 μg/ml (8). No prior reports of testing susceptibility to the echinocandins were found. It was predicted that Rhodotorula species would be more resistant than Candida species to caspofungin, as the echinocandins are typically not effective for another heterobasidiomycete, Cryptococcus neoformans (7). Awareness of poor activity against heterobasidiomycetes is necessary, as echinocandins are often instituted after yeasts are recovered from blood cultures and prior to their identification.

Treatment of Rhodotorula infection involved the removal of CVCs and, generally, 14 days of amphotericin B or fluconazole therapy. Specimens from two patients treated successfully with fluconazole had Rhodotorula pilimanae (now taxonomically identified as R. mucilaginosa) and Rhodotorula glutinis isolates, which grew poorly at 35°C in vitro. The MIC of fluconazole used in these cases was high, 32 μg/ml, but successful treatment with an antifungal agent demonstrating an elevated MIC could be attributed to catheter removal alone or to decreased hardiness of the temperature-sensitive phenotype. One nonimmunocompromised patient had clinical failure with fluconazole but cleared fungemia with amphotericin B and catheter removal. One patient cleared fungemia with two doses of amphotericin B alone.

Clinical data include previous reports of successful therapies for Rhodotorula BSIs, including recovery of neutropenia (11), catheter removal (10, 15), amphotericin B (14), or combinations thereof (1, 4). The largest reported series of Rhodotorula fungemia demonstrated favorable outcomes in all patients with either catheter removal or amphotericin B treatment (16, 10).

Conclusions.

Infections from Rhodotorula species are uncommon but are observed in hosts with CVCs and/or immunosuppression. Amphotericin B preparations, in addition to catheter removal, are acceptable therapies for Rhodotorula infection, with excellent in vitro activity and reports of successful use. Flucytosine possesses excellent activity in vitro. Based on our in vitro data, narrow spectrum azoles are not appropriate therapy. Further studies are needed to determine the role of extended-spectrum azoles, given the wide spectrum of activity against Rhodotorula species. Echinocandins should not be considered appropriate therapy for Rhodotorula species.