First Case of Bloodstream Infection Due to Candida magnoliae in a Chinese Oncological Patient

Abstract

We report a case of fungemia caused by Candida magnoliae, a yeast never associated with human disease. The infection occurred in a 42-year-old Chinese patient with gastric cancer complicated by peritoneal carcinosis. Multiple blood cultures were positive for yeast; the species was well identified with biochemical and molecular methods. The phylogenetic analysis showed a close relationship of C. magnoliae to Candida krusei.

CASE REPORT

A 42-year-old male patient of Chinese origin who had previously undergone surgery for gastric cancer, which was subsequently complicated by peritoneal carcinosis, presented to the Verona University Hospital casualty department for anorexia, vomiting, asthenia, and weight loss. Physical examination revealed ascites and clinical signs of intestinal subocclusion. Nineteen months earlier, the patient had undergone a subtotal distal gastrectomy and lymphadenectomy for a diffuse-type gastric carcinoma with positivity for Helicobacter pylori (pT3-pN9/34-M0). A central venous catheter had been positioned, after which the patient had been treated with two cycles of an ECF chemotherapy protocol (epirubicin, cisplatin, and fluorouracil) followed by combined radiotherapy (45 Gy in 25 sessions) and chemotherapy (four cycles of 5-fluorouracil and folinic acid). Four months later, peritoneal carcinosis was detected, so the patient was treated with a further eight cycles of chemotherapy consisting of oxaliplatin, 5-fluorouracil, and folinic acid; the last cycle was administered 2 months prior to presentation to the casualty department.

Physical examination showed a cachectic, apyretic patient with ascites and clinical signs of intestinal subocclusion. The patient's blood tests gave the following results: hematocrit, 28.9%; hemoglobin, 9.7 g/dl; red blood cell count, 3.2E12/liter; white blood cell count, 5.49E09/liter with 4,900 neutrophils/mm³ and 810 lymphocytes/mm³; K⁺, 3.1 mmol/liter; total protein, 63.4 g/liter with 33.9 g/liter albumin; and erythrocyte sedimentation rate, 40 mm/h. Chest X-ray findings were normal. Computed tomography confirmed peritoneal carcinosis.

During the patient's stay in hospital, the patient was treated with 500 mg levofloxacin intravenously (i.v.) once daily for 12 days for a suspected urinary tract infection, 1 g cefotaxime i.v. three times daily for 7 days, later replaced by 1 g meropenem i.v. three times daily for 5 days because of suspected bacterial peritonitis, and mouth-cleansing nystatin for 15 days until discharge. During the patient's hospitalization, the patient underwent evacuative paracentesis five times, with the emission of about 2,200 ml of serum-hematic fluid each time. No microbiological culture was done.

On day 21, the patient was febrile (38.2°C), and six blood cultures (BactAlert FAN; bioMérieux) were performed over a 4-day period. After a 48-h incubation, four blood cultures became positive for a yeast. The patient was given 1 mg/kg of body weight amphotericin i.v. for 3 days, which was then replaced by 200 mg fluconazole i.v. for 14 days. After about 2 days, the patient's temperature reverted to normal, and the patient was discharged 10 days later. The patient died 1 month later as a result of cancer progression and deterioration.

The morphological, physiological, and biochemical characteristics of the four strains isolated from the patient's blood on days 23 and 25, respectively, were examined by standard methods commonly used in yeast taxonomy (2). An ID 32C panel (bioMérieux) and a YST card for Vitek 2 (bioMérieux) were used for identification of the yeast strains, according to the manufacturer's instructions. The yeast isolates from the patient's blood cultures grew after 24 to 48 h of incubation at 37°C in the Bact/Alert System (bioMérieux). Colonies developed on Sabouraud's dextrose agar over 7 days of incubation at 25°C and were white to cream colored, soft, and smooth. The germ tube test was negative. No hyphae or pseudohyphae were observed. Microscopic morphology on a Dalmau plate (Fig. 1) showed globose to oval cells, single or budding, with thick-walled chlamydospores. No pseudomycelium was seen. The yeast failed to grow on cycloheximide-containing medium and was negative in the urease test and positive in the nitrate assimilation test. It fermented glucose and sucrose only. All strains gave excellent identification for Candida magnoliae with the use of ID 32C (profile 4020110211; 99.4% agreement) and the YST card for Vitek 2. One of these strains, C. magnoliae SL040806, was selected for further study.

FIG. 1.

Morphological features of Candida magnoliae on cornmeal agar at 7 days of growth at ambient temperature: globose to oval cells, single or budding. No pseudomycelium is seen.

To confirm the biochemical identification, genomic DNA was isolated from fungal colonies grown on Sabouraud agar plates, as described previously (9), and sequence analysis of the 18S ribosomal DNA (a highly conserved region in Candida species) was performed as described by Suzuki et al. (15); the sequence alignment was done by using the CLUSTAL W program (1). Comparative analysis with the C. magnoliae sequence deposited in GenBank (accession number AB018145) revealed strong identity (99.4%).

The phylogenetic relationship of C. magnoliae to other Candida spp. was investigated by sequencing the internal transcribed spacer (ITS) region located between the 18S and 26S rRNA genes and subdivided into the ITS1 region between the 18S and 5.8S rRNA genes and the ITS2 region between the 5.8S and 26S rRNA genes (4, 5). This region is more divergent and distinctive than the highly conserved rRNA genes for Candida species (10, 16) and has never been examined in C. magnoliae. The primers and conditions used to amplify the ITS region were as described previously (11, 12, 17). The ITS sequences of the C. magnoliae SL040806 strain and those of the reference strains obtained from GenBank were aligned using the CLUSTAL W program (1), and a phylogenetic tree was generated (Fig. 2). Phylogenetic analysis was performed according to Suzuki et al. (15), using the neighbor-joining method and the two-parameter model; confidence values for individual branches were determined by bootstrap analysis. The accession numbers of the reference sequences are indicated in the tree shown in Fig. 2.

FIG. 2.

Phylogenetic tree generated from neighbor-joining analysis of the ITS region sequences and depicting the relationships of Candida magnoliae to closely related yeast taxa. Reference sequences were from the type strains of the species and were retrieved from GenBank under the accession numbers indicated.

Antifungal susceptibility testing was initially performed by the broth microdilution method according to the guidelines outlined in CLSI (formerly NCCLS) document M27-A2 (13). RPMI medium and an inoculum size of 5 × 10³ CFU were used. Since no growth was detected under the standard incubation conditions, the susceptibility test was repeated using incubation at 35°C in air, supplemented or not with 10% CO₂, until growth occurred. Under these conditions, the yeast showed apparent turbidity in 4 days' incubation with CO₂ and in 14 days in air. A change in turbidity equal to or greater than 90% compared to the results for drug-free controls was used to establish the MIC breakpoints under both conditions. The results are summarized in Table 1. Fluconazole resistance was detected both after 14 days' incubation in ambient air and after 4 days in air supplemented with 10% CO₂. Candida parapsilosis ATCC 22019 and Candida krusei ATCC 6258, used as quality controls, gave MICs within the expected range in both conditions, thus excluding the possibility that drug inactivation may have given false resistance results.

TABLE 1.

Antifungal susceptibility testing^a

Drug	MIC (μg/ml) for Candida magnoliae SL040806 under condition:		MIC (μg/ml) for Candida parapsilosis ATCC 22019 under condition:		MIC (μg/ml) for Candida krusei ATCC 6258 under condition:
	A	B	A	B	A	B
Amphotericin	0.016	0.125	0.5	1	1	1
Flucytosine	<0.03	0.25	0.5	0.5	16	16
Fluconazole	>256	>256	4	4-8	128	128
Itraconazole	0.06	0.06	0.5	0.5	0.25	0.25
Ketoconazole	0.03	0.03	0.125	0.25	0.5	0.5
Voriconazole	0.25	0.5	0.06	0.125	0.5	1

^aA, incubation at 37°C for 4 days in air supplemented with 10% CO₂; B, incubation at 37°C for 14 days in air.

To the best of our knowledge, Candida magnoliae has been implicated in just one case of human disease (tenosynovitis in an immunocompetent child) (7). Our report describing a bloodstream infection by C. magnoliae in a terminal oncological patient is the first to indicate the role of this yeast also in candidemia.

The incidence of bloodstream infections caused by Candida spp. has increased progressively over the past two decades (6, 14), and among the Candida species identified in bloodstream infections to date, Candida albicans ranks first, though recent studies show a shift towards an increasing prevalence of non-C. albicans species. This change has been attributed in part to the widespread use of antifungal agents and to the increasing number of immunocompromised hosts; however, it may also reflect the increasing awareness of non-C. albicans Candida species as important opportunistic pathogens (3). It has also become apparent that species once considered to be of only industrial importance or innocuous inhabitants of the environment are capable of attacking the human host. These organisms may vary greatly in their susceptibility to the current antifungal agents, thus escaping the effects of empirical treatments.

Candida magnoliae is a yeast isolated from magnolia flowers and from the bumblebee gut. It is also an industrially important yeast with substantial erythritol-producing ability. Erythritol has been used as a functional sugar substitute for various foods because it is noncarcinogenic, a low-calorie sweetener, and safe for diabetics (18). It is also used in biotechnology to produce mannitol from glucose (8).

A worrying feature of the infection caused by C. magnoliae was the very slow growth of the yeast under the conditions recommended by the CLSI guidelines for antifungal susceptibility tests. Apparently, the yeast was resistant to fluconazole under the modified conditions used by us, and the possibility that this resistance was innate was suggested by the strict phylogenetic relationship we found to C. krusei (Fig. 2). We do not know whether the resistance detected under these conditions was clinically significant, since the patient was first treated with amphotericin, to which the yeast was “susceptible” under both conditions. Since the shift to fluconazole therapy occurred after 3 days of amphotericin use, the drop in fever could have been linked to the activity of the first drug used.

Considering the patient's Chinese origins, we formulated different speculative hypotheses as to how the patient may have contracted this type of Candida: (i) the patient may have consumed products such as yoghurt, cheese, or honey which contain yeasts such as Candida magnoliae that, in view of his condition, may have entered the bloodstream; (ii) the patient may have taken a magnolia-based medicine, used in Chinese medicine to treat gastrointestinal disorders; or (iii) the patient may have used sugar-free products with traces of this yeast. Language difficulties made it impossible to investigate any of these hypotheses in depth.

Clinical microbiology laboratories and physicians will often be involved in identifying and treating infections with unusual yeasts not previously considered to be pathogens. The great variability of these species in terms of their susceptibility to the current antifungal agents entails further difficulty with regard to the empirical treatment of these potential infections. Candida magnoliae should be added to the long list of yeasts capable of causing bloodstream infections in immunocompromised or critically ill patients.

Nucleotide sequence accession number.

The ITS region sequences of the strain C. magnoliae SL040806 have been deposited in GenBank under accession no. AM408497.