Candida fermentati as a Cause of Persistent Fungemia in a Preterm Neonate Successfully Treated by Combination Therapy with Amphotericin B and Caspofungin

Abstract

A case of persistent candidemia in a preterm neonate caused by Candida fermentati, identified by sequencing of the internally transcribed spacer region of ribosomal DNA (rDNA), is described. The neonate was treated for 30 days by combination therapy with amphotericin B (AmBisome) and caspofungin with a successful outcome, and no drug-related side effects were observed.

CASE REPORT

A 27-week-preterm male, weighing 1,040 g, was delivered to a Kuwaiti mother by emergency longitudinal cesarean section in a surgical intensive care unit. On day 0, the baby was diagnosed to have hyaline membrane disease and patent ductus arteriosus with Apgar scores of 7 and 9. Soon after delivery, an umbilical artery catheter (UAC), an umbilical venous catheter (UVC), and a peripheral cannula were placed to provide fluids and medication and the baby was also intubated. On day 5, he was empirically started on ampicillin and amikacin. On day 12, he developed signs and symptoms of sepsis requiring a change in treatment to cloxacillin and amikacin. Since his septic condition continued, the treatment was changed to vancomycin. On day 27, the blood culture grew a yeast (Kw3414/13), which was identified as Candida famata by the use of a Vitek 2 yeast identification (YST ID) card (bioMérieux, Marcy I'Étoile, France). While cultures of UVC and UAC tips were negative for growth, a culture from a long-line tip yielded yeast which was also identified as C. famata. The antifungal drug susceptibility was determined by Etest, and the isolate appeared susceptible to amphotericin B (Table 1). The baby was started on amphotericin B at 1.5 mg/kg of body weight. A blood culture repeated after 2 days of treatment again yielded C. famata, leading to a change of treatment to a lipid formulation of amphotericin B (AmBisome) (8.5 mg once daily). Despite treatment with amphotericin B for 5 days, the blood culture remained positive for C. famata, prompting addition of caspofungin (2.5 mg/kg). Although additional blood cultures performed on day 50, day 53, and day 59 yielded C. famata (identified by the use of a Vitek 2 yeast identification system), there was symptomatic improvement in the condition of the baby. A blood culture repeated on day 66 (after 30 days of combination therapy with amphotericin B and caspofungin) became negative for the yeast. There was no recurrence of infection during 3 weeks of follow-up, and the baby was subsequently discharged in healthy condition.

TABLE 1.

Antifungal susceptibility of C. fermentati strains isolated from blood samples

Strain or parameter	Date of isolation	MIC by Etest (μg/ml)a
		AP	FL	VO	POS	IT	FC	CS	MYC
Kw3414/13	30 September 2013	0.002	3	0.125	0.19	3	0.003	0.125	0.38
Kw3434/13	2 October 2013	0.047	3	0.125	0.19	4	0.006	0.19	0.25
Kw3476/13	7 October 2013	0.064	4	0.125	0.19	4	0.004	0.38	0.38
KW3556/13	21 October 2013	0.094	4	0.38	0.19	3	0.003	0.19	0.38
Kw3593/13	24 October 2013	0.094	3	0.19	0.25	4	0.012	0.38	0.25
Kw3644/13	29 October 2013	0.125	4	0.38	0.25	4	0.016	0.75	0.75
Geometric mean of MICs		0.043	3.46	0.194	0.208	3.634	0.005	0.280	0.370

^aAbbreviations: AP, amphotericin B; FL, fluconazole; VO, voriconazole; IT, itraconazole; POS, posaconazole; FC, flucytosine; CS, caspofungin; MYC, micafungin.

During 6 weeks of follow-up, 6 blood culture isolates were obtained and were characterized. All isolates formed whitish colonies on Sabouraud dextrose agar which became slightly tanned on aging. Microscopic examination revealed ovoid to elongated yeast cells of variable sizes (2 to 5.2 by 2.2 to 5.6 μm) occurring singly, in pairs, or in short chains. A Dalmau slide culture on cornmeal agar at 30°C showed well-branched pseudohyphae with clusters of budding yeast cells (Fig. 1). True hyphae were lacking. Growth on ascospore agar and malt extract agar, incubated at 25°C for 30 days, did not show ascospores. Based on the results of carbohydrate assimilation profiles obtained by ID32C and a Vitek 2 yeast identification system, the isolates were identified as C. famata. For molecular identification, the genomic DNA from the isolates was prepared as described previously (1). The internally transcribed spacer (ITS) region of ribosomal DNA (rDNA) was amplified and sequenced as described previously (2). Molecular fingerprinting of the isolates was carried out by PCR sequencing of the intergenic spacer region between the 28S rRNA and 5S rRNA genes by using IGS1F (5′-CGGAGTATTGTAAGCAGTAGA-3′) and IGS1R (5′-TAGTGGGAGACCATACGCGAA-3′) as PCR amplification and sequencing primers, essentially as described previously (3). PCR fingerprinting of the isolates was also carried out with microsatellite-based (GACA₄, 5′-GACAGACAGACAGACA-3′) and minisatellite-based (M13-MIN, 5′-GAGGGTGGCGGTTCT-3′) primers as described previously (4). The ITS region sequences of our isolates were identical and showed complete (100%) identity with the corresponding sequences from reference Candida fermentati strains CBS 9966 (type strain) and CBS 2022 but showed 4 nucleotide differences from the sequence from Candida guilliermondii CBS 2030, thus establishing their identity as C. fermentati (5). The IGS region sequences of all the isolates were identical. PCR fingerprinting carried out with microsatellite-based and minisatellite-based primers also yielded identical amplified DNA fragment patterns from all the 6 isolates, indicating clonality.

FIG 1.

Dalmau slide culture on cornmeal agar showing branched pseudohyphae with budding yeast cells.

A case of persistent fungemia caused by C. fermentati in a preterm neonate is described. All six C. fermentati isolates showed antifungal susceptibility profile consistent with the published studies (6–9), and there was no indication of acquisition of resistance during treatment (Table 1). A noteworthy feature of our case is the successful use of caspofungin in combination with amphotericin B for an extended period of 30 days without any evidence of side effects or toxicity. There appears to be no previous report of the combined use of these two drugs in the management of candidemia in preterm neonates. Presently, there is no recommended dose for caspofungin for the treatment of invasive candidiasis in neonates (10). However, based on limited clinical data, a dosage of 25 mg/m²/day has been suggested, which is comparable to the dosages corresponding to the results of area under curve/pharmacokinetic studies carried out in adults (11). In some studies, caspofungin has been used in neonates who were either refractory to or intolerant of amphotericin B therapy without causing any adverse effects (12, 13).

Candida fermentati (teleomorph Meyerozyma caribbica) is another pathogenic species within the C. guilliermondii complex, representing about 9% of the isolates in this complex in the global collection originating from different geographic regions (6). There are only two accepted species in the genus Myerozyma, M. caribbica and M. guilliermondii, although some additional closely related taxa have been proposed upon further molecular analysis (14). These species are estimated to cause about 1% to 3% of the total number of cases of candidemia (6, 7). Candida fermentati was first described in 2005 by Vaughan-Martini et al. as an anamorphic species of Pichia carribica under the Pichia guilliermondii complex (15). It was later placed in the ascosporic genus Meyerozyma, as M. caribbica, along with M. guilliermondii (14). Members of the C. guilliermondii complex are indistinguishable by conventional phenotypic methods but appear genetically heterogeneous (16–18). Although C. guilliermondii is widely distributed in the environment and has been isolated from a variety of saprophytic sources, it plays an infrequent role in human diseases (19). So far, C. fermentati and C. guilliermondii are the known human pathogens in this complex. Some recent reports suggest that the incidence of C. guilliermondii or C. fermentati fungemia is highest in Latin America (7, 20), China (21, 22), and Taiwan (16), particularly in pediatric-care age groups (20–23). In a recent Chinese study, 39 of 238 (16.5%) patients had candidemia due to C. guilliermondii (22). A multivariate analysis of the risk factors in these patients revealed that preterm infants with low birth weight were at the higher risk of candidemia due to C. guilliermondii strains than non-C. guilliermondii strains (51.3% [20/39] versus 8.5% [17/199], respectively; P = 0.001). Other risk factors included intravenous nutrition and surgery (22). In an earlier series of 9 neonatal cases of Candida septicemia from China, 6 were caused by C. guilliermondii (21). The reasons for the increased rate of occurrence of C. guilliermondii candidemia in these two studies need further epidemiological investigations.

C. fermentati/C. guilliermondii complex isolates are known to exhibit reduced in vitro susceptibility to azoles and echinocandins (7–9, 19, 24, 25), which may have therapeutic implications. In addition, the possibility of a gradual increase in their incidence as a result of selection pressure also exists. C. guilliermondii complex isolates resistant to amphotericin B and azoles have been isolated from clinical specimens (26, 27). In one study conducted under the auspices of the ARTEMIS DISK Antifungal Susceptibility Program, 10.8% and 4.9% isolates of C. guilliermondii were resistant to fluconazole and voriconazole, respectively (7). Likewise, C. guilliermondii strains are found to be 2-fold to 16-fold less susceptible than other Candida species (excluding C. parapsilosis) to caspofungin, anidulafungin, and micafungin (7, 28). In a multicenter study, Pfaller et al. (9) examined 234 C. guilliermondii strains to determine epidemiologic cutoff values (ECVs) for anidulafungin and micafungin. MICs for ECVs for anidulafungin and micafungin covering 97.5% of the strains were 8 and 2 μg/ml, respectively. Similarly, based on wild-type MIC distributions, the ECVs for C. guilliermondii (n = 373) for fluconazole, posaconazole, and voriconazole were 8, 0.5, and 0.12 μg/ml, respectively, in another study (10). Interstrain differences in susceptibility to azoles depending upon the source of isolation have also been noted among C. guilliermondii isolates (9). For example, 85% of bloodstream isolates of C. guilliermondii were susceptible to fluconazole, whereas only 67.7% from skin and soft tissues were susceptible. In contrast, 93.4% of isolates of C. guilliermondii originating from blood, skin, and soft tissues were susceptible to voriconazole whereas only 80.4% of isolates from the urinary tract were susceptible to this antifungal drug (7).

Here, it is pertinent to mention that most of the antifungal susceptibility data on C. fermentati reported in the literature are derived from isolates representing the C. guilliermondii complex (29). This is because the available phenotypic methods alone are insufficient to provide unequivocal identification of the species included in the complex (30). The importance of correct species identification is underscored by species-specific interpretive criteria. Limited in vitro susceptibility data suggest that there are differences between the antifungal susceptibility profiles of C. fermentati and C. guilliermondii. While C. guilliermondii isolates tend to show higher micafungin MICs (0.57 versus 0.35 μg/ml, P = 0.0001), C. fermentati isolates, on the other hand, have higher fluconazole (6.13 versus 3.17 μg/ml, P = 0.079) and amphotericin B (0.31 versus 0.18 μg/ml, P = 0.088) MICs (6). To what extent these differences in antifungal susceptibilities between the two species impact therapeutic outcome is unclear.

Yeast species with similar phenotypic characteristics have always posed identification problems (5, 30). Conventional identification methods often yield erroneous species identification. C. famata is often wrongly identified by conventional identification methods such the use of Vitek 2 (5, 6, 16, 18, 30, 31). Matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) platforms cannot be used for rapid identification of culture isolates since C. fermentati is not currently included in their database. Hence, any C. famata isolation from clinical specimens should be confirmed by sequencing of rDNA.

In conclusion, a case of persistent candidemia due to C. fermentati, identified by sequencing of the ITS region of rDNA, is described. The neonate was treated with amphotericin B and caspofungin combination therapy for 30 days with no drug-related side effects. To the best of our knowledge, this is the first report of the successful use of amphotericin B and caspofungin in combination in a preterm neonate for the treatment of persistent candidemia. Additionally, the report highlights the use of molecular methods in the identification of rare or cryptic Candida species causing human infections.

Nucleotide sequence accession numbers.

Nucleotide sequences determined in this work have been deposited in the EMBL database under accession no. HG970745, HG970748, LK392385 to LK392391, and LN651284 to LN651286.