European Journal of Clinical Microbiology and Infectious Diseases

, Volume 25, Issue 3, pp 193–196

Rhodotorula mucilaginosa outbreak in neonatal intensive care unit: microbiological features, clinical presentation, and analysis of related variables

Authors

    • Neonatal Intensive Care UnitV. Fazzi Regional Hospital
  • M. L. Faneschi
    • Laboratory Medicine Unit, Microbiology SectionV. Fazzi Regional Hospital
  • E. Manso
    • Laboratory Medicine Unit, Microbiology Section, Umberto I–G. M. Lancisi–G. Salesi United HospitalsUniversity of Ancona
  • M. Pizzolante
    • Laboratory Medicine Unit, Microbiology SectionV. Fazzi Regional Hospital
  • A. Rizzo
    • Laboratory Medicine Unit, Microbiology SectionV. Fazzi Regional Hospital
  • A. Sticchi Damiani
    • Laboratory Medicine Unit, Microbiology SectionV. Fazzi Regional Hospital
  • R. Longo
    • Neonatal Intensive Care UnitV. Fazzi Regional Hospital
Concise Article

DOI: 10.1007/s10096-006-0114-2

Cite this article as:
Perniola, R., Faneschi, M.L., Manso, E. et al. Eur J Clin Microbiol Infect Dis (2006) 25: 193. doi:10.1007/s10096-006-0114-2

Abstract

Reported here are the features of a Rhodotorula mucilaginosa outbreak that occurred in a neonatal intensive care unit. Over a period of 19 days, clinical and laboratory signs of sepsis appeared in four premature infants carrying indwelling vascular catheters. After bloodstream infection with R. mucilaginosa was ascertained, the patients underwent amphotericin B therapy and recovered completely. In a retrospective case-control study, the variables displaying a statistical difference between case and control-group neonates were birth weight, gestational age, duration of parenteral nutrition, duration of antibiotic therapy and prophylactic administration of fluconazole. To our knowledge, this is the first reported outbreak caused by yeasts of the Rhodotorula genus.

Introduction

Rhodotorula is an anamorphic yeast genus belonging to the phylum Basidiomycota [1]; although considered saprophytes or contaminants [2, 3], in the last few years some Rhodotorula species, especially Rhodotorula mucilaginosa, have emerged as opportunistic pathogens [2, 47]. Presented here is an analysis of the first documented outbreak of R. mucilaginosa in a neonatal intensive care unit (NICU). In the outbreak, four premature infants were infected over a period of 19 days, and therapy with amphotericin B led to elimination of the yeast in all cases.

Patients and methods

Four cases of bloodstream infection caused by R. mucilaginosa were registered in January and February 2003 at the NICU of the V. Fazzi Regional Hospital in Lecce, Italy. The first blood culture positive for R. mucilaginosa was obtained on 23 January, and the last on 11 February. For blood cultures, 1–4 ml of blood was placed in paediatric BacT/Alert PF bottles (bioMérieux, Hazelwood, MO, USA), which were monitored by BacT/Alert 3D (Organon Teknika, Durham, NC, USA).

Microscopic examination of wet preparations from growth-positive bottles showed budding yeasts with narrow bud scars without hyphae or pseudohyphae. After incubation for 48 h at 35°C, subcultures on Sabouraud glucose agar grew mucoid orange colonies. The Vitek Yeast Biochemical Card (bioMérieux) identified Rhodotorula glutinis/R. mucilaginosa with 90% reliability: the definitive recognition of R. mucilaginosa relied on the absence of nitrate assimilation. Identification was confirmed with 99.2% reliability using API ID 32C (bioMérieux): ribose, glycerol, galactose, sucrose, l-arabinose, raffinose and glucose were assimilated, while the other substrates, including trehalose, were not. Trehalose is assimilated in 85% of R. mucilaginosa strains: in addition to the spatial-temporal correlation, failure to assimilate trehalose was considered to confirm the identity of the four isolates. R. mucilaginosa had never been previously isolated in this NICU.

Susceptibility to antifungal agents was assayed using the Sensititre YeastOne Colorimetric Antifungal Panel (TREK, East Grinstead, UK), and after 72 h of incubation at 35°C, the following MICs were obtained: fluconazole >256 mg/l, itraconazole 2 mg/l, voriconazole 2 mg/l, ketoconazole 0.25 mg/l, amphotericin B 0.25 mg/l, and flucytosine 0.125 mg/l. Intravenous therapy with liposomal amphotericin B was initiated 6–7 days after blood culture. In the first infected neonate, the administration of amphotericin B was delayed because of initial uncertainty about the real virulence of the yeast; in the meantime, more exhaustive information was acquired about R. mucilaginosa infections. Significantly, a second blood culture positive for R. mucilaginosa (tested by Vitek alone) was obtained for this neonate before antifungal therapy was started. Blood cultures performed at the end of the antifungal therapy were all negative.

All four neonates had had a vascular central line (umbilical venous catheter, central venous catheter, or both) since birth, but early removal and substitution following confirmation of R. mucilaginosa sepsis was possible for only two of them. Three patients were on fluconazole prophylaxis according to the following schedule: an intravenous dose of 3 mg/kg body weight every third day for the first 2 weeks, every other day during weeks 3 and 4, and daily during the next 2 weeks [8]. Prophylactic administration of fluconazole to prevent invasive fungal infections in premature infants of gestational age <32 weeks carrying vascular central lines was used in this NICU from the second semester of 2002 throughout 2003; however, this practice was abandoned thereafter due to the occurrence of “breakthrough” fungaemia and the emergence of Candida albicans strains and other fungi not susceptible to fluconazole.

In an attempt to find the source of the outbreak, cultures were obtained from the hands of personnel, parenteral nutrition bags and various objects (e.g., incubators, sinks, basins, surgical instruments and their recipients, topical medications and open vials of pharmaceuticals), but no colonization or contamination was demonstrated. On removal of the vascular catheters, cultures of the catheter tips were performed, but these were also negative. After the fourth case of R. mucilaginosa sepsis was detected, the NICU was thoroughly cleaned and disinfected. At the time of writing, no further R. mucilaginosa isolates had been recovered in the NICU.

To test the ability of the infected neonates to recognize the yeast antigenic stimulus, some days after the third case of R. mucilaginosa sepsis was ascertained all three patients infected by that time underwent a lymphocyte proliferation test. Following the reconstitution of lyophilized culture mediums, 250 μl of heparinized blood was incubated with 200 μl of R. mucilaginosa concentrated to 1×106/μl. After erythrocyte lysis and supernatant discarding, the pellet was suspended in 1 ml of propidium iodide staining solution (Ylem, Avezzano, Italy). Lymphocyte response was expressed as the percentage of S-phase cells as detected by flow cytometry. Six healthy persons (two neonates and four adults) also participated in the test.

Regarding follow-up of the infected neonates, no brain lesions were detected during subsequent ultrasound examinations; similarly, the electroencephalograms were normal. At the age of about 2.5 years the children showed no signs of neurological impairment.

The medical records of the four infected neonates were examined in order to determine the clinical and laboratory characteristics of the R. mucilaginosa sepsis. Subsequently, a retrospective study was performed to detect the variables statistically related to R. mucilaginosa sepsis. For this analysis, a control group was created consisting of 13 non-case premature infants (gestational age ≤35 weeks), hospitalized in the NICU for at least 72 h between 20 January (i.e., 72 h before the first blood culture positive for R. mucilaginosa) and 16 February (i.e., the date of disinfection of the NICU). Altogether, the control-group neonates underwent 21 culture tests from various sources (nine blood cultures, seven cultures of indwelling catheter tips, three urine cultures, and one culture each of tracheal tube and gastric aspirate); none of these was positive for R. mucilaginosa. The length of hospital stay and the other clinical and laboratory records were calculated at the date of the positive blood culture for R. mucilaginosa in the case neonates, and at the date of discharge in the non-case neonates.

The continuous variables compared using the Mann–Whitney test were birth weight, gestational age, and duration of hospitalization, incubator stay, mechanical ventilation, vascular central line, parenteral nutrition, enteral feeding and antibiotic therapy. For some of these variables, a further comparison was obtained by calculating the mean duration of exposure per 100 patient-days. Two dichotomic variables (bacterial bloodstream infection and fluconazole prophylaxis) were compared using Fisher’s exact test (Prism 3.0; GraphPad, San Jose, CA, USA). A two-tailed p value of <0.05 was considered statistically significant.

Results and discussion

As shown in Table 1, the prominent clinical signs observed in the neonates infected with R. mucilaginosa were lethargy, tachycardia and evidence of respiratory distress. Three patients showed mild abdominal distension and transitory feeding intolerance, while severe hypotension did not occur. All four neonates had thrombocytopenia; two also had neutropenia, and all but one patient demonstrated an increased C-reactive protein level. Previous bacterial bloodstream infection had occurred in three cases. The results of the lymphocyte proliferation test with R. mucilaginosa did not show any difference between case neonates and healthy persons (mean percentage of S-phase cells, 17.3±5.5 vs 18.8±6.0; p not significant). Comparison of continuous variables (Table 2) shows that birth weight, gestational age and duration of parenteral nutrition and antibiotic therapy were significantly different between case and control-group neonates. The duration of enteral feeding was not different in its absolute value, but when the mean value per 100 patient-days was considered, it seemed to exert a protective role. Finally, Fisher’s exact test revealed a statistically significant difference for the fraction of neonates who received fluconazole prophylaxis.
Table 1

Birth and Rhodotorula mucilaginosa sepsis-related characteristics of the case neonates (ordinal numbers refer to the patients’ age in days)

Characteristic

Patient number/sex

1/F

2/M

3/F

4/F

Birth weight (g)

1,350

1,340

1,750

680

Gestational age (weeks)

31

28

31

30

Previous bacterial BSI

Yes

Yes

Yes

No

Isolate

Escherichia coli

Staphylococcus epidermidis

Escherichia coli

Fluconazole prophylaxis

Yes

No

Yes

Yes

Rhodotorula mucilaginosa culture

Jan 23 (21st), Feb 3 (32nd)

Jan 27 (8th)

Jan 28 (23rd)

Feb 11 (21st)

Body temperature (°C)

Normal

Normal

Abnormal (36.2)

Normal

Lethargy

Yes

Yes

Yes

Yes

Tachycardia

Yes

No

Yes

Yes

Respiratory distress

Yes

Yes

No

Yes

Gastrointestinal signs

Yes

No

Yes

Yes

Severe hypotension

No

No

No

No

Hyperglycemia (mmol/l)

No

Yes (12.2)

No

No

Metabolic acidosis (pH)

Yes (7.22)

Yes (7.23)

No

Yes (7.23)

Neutropenia (cells/μl)

Yes (620)

No

No

Yes (470)

Thrombocytopenia (cells/μl)

Yes (47,000)

Yes (91,000)

Yes (1,000)

Yes (33,000)

C-reactive protein (mg/dl)

Increased (20.0)

Normal

Increased (3.5)

Increased (47.1)

Vascular central line removal

Late (45th)

Early (8th)

Late (38th)

Early (24th)

Amphotericin B therapy

Yes (32nd–45th)

Yes (15th–26th)

Yes (29th–37th)

Yes (28th–41st)

F female, M male, BSI bloodstream infection

Table 2

Comparison of continuous variables between case and control-group neonates. Data shown represent mean values and standard deviations; units are days unless otherwise specified

Variable

Absolute value

Value per 100 patient-days

Cases (n=4)

Controls (n=13)

p value

Cases (n=4)

Controls (n=13)

p value

Birth weight (g)

1,280±443

2,160±476

0.009

   

Gestational age (weeks)

30.4±1.2

34.3±1.3

0.004

   

Hospital stay

18.4±6.7

15.0±11.0

NS

   

Incubator stay

18.4±6.7

10.1±9.0

NS

   

Mechanical ventilation

1.7±1.3

0.5±1.0

NS

   

Vascular central linea

19.2±7.2

6.0±9.9

NS

104.3±7.1

25.0±35.6

0.004

Parenteral nutrition

17.2±6.9

3.9±6.8

0.031

92.0±6.7

17.4±26.7

0.005

Enteral feeding

11.2±6.1

12.3±8.8

NS

58.7±16.6

84.3±11.3

0.015

Antibiotic therapyb

26.6±9.8

9.2±12.4

0.031

151.3±38.4

52.0±63.7

0.036

Bacterial BSI (n)

3

2

NS

   

Fluconazole prophylaxis (n)

3

1

0.022

   

NS not significant, BSI bloodstream infection

aCalculated for each neonate as the sum of the umbilical and central venous catheter duration

bCalculated for each neonate as the sum of the duration of therapy with β-lactam, aminoglycoside and glycopeptide antibiotics

The four cases reported here demonstrate that R. mucilaginosa is able to cause nosocomial outbreaks. Although the first reports of Rhodotorula sepsis date back to the 1960s [911], yeasts belonging to this genus were not considered emerging pathogens until the end of the 1980s [4]. Cases of R. mucilaginosa infection in infants under 1 year of age are exceptionally rare [10]. Moreover, no outbreak of R. mucilaginosa has been reported previously: in our analysis of the outbreak, both the spatial-temporal relationships among the cases, and the phenotypic patterns of the four isolates support their common identity. This is reinforced by the fact that R. mucilaginosa had never previously been isolated in the NICU.

Evaluation of the clinical signs and variables statistically related to R. mucilaginosa sepsis indicates that this disease shares the same features as other typical neonatal fungemias [12]. However, since no mortality and no neurological sequelae were detected in the follow-up of the infected neonates, on the whole the virulence displayed by the yeast was fairly low. On the other hand, the very small size of the matched samples necessarily limited the scope of the statistical analysis; specifically, it was impossible to perform more advanced tests such as multivariate analysis adjusting for birth weight and gestational age. Attempts to detect the source of the outbreak were unsuccessful; thus, the way in which R. mucilaginosa spread remains unexplained. The literature indicates that R. mucilaginosa fungemia is generally observed in patients undergoing prolonged intravenous therapy or parenteral nutrition via indwelling catheters [47]. Concomitant immunosuppression of the patients seems to be indispensable to the pathogenicity of the yeast. In this sense, premature infants are at high risk of sepsis because of the immaturity of their immune systems, despite their ability to recognize the antigenic stimulus (as shown by the results of the lymphocyte proliferation test); the risk is increased by the greater number of invasive procedures these patients undergo [13]. Fungal colonization and proliferation are also favored by prolonged antibiotic therapy. In the cases reported here, an additional risk factor was presumably represented by antifungal prophylaxis: since most R. mucilaginosa strains are intrinsically resistant to fluconazole [14], its administration could have selected for this yeast compared to other fungi.

There is some debate concerning the countermeasures to be adopted if R. mucilaginosa sepsis occurs. Some authors suggest removing the indwelling catheter and administering antifungal therapy [6, 7]; other authors, believing this yeast to have low pathogenicity, consider one of the above options alone to be sufficient [2, 5]. When dealing with this matter, the difficulties, risks and costs of removing and replacing a vascular central line in a premature infant have to be taken into account [15]. Among our case patients, it is important to note that the neonate who was not initially given amphotericin B, and in whom the vascular catheter was not removed early, had a second blood culture positive for R. mucilaginosa.

A final remark should be made on the susceptibility of R. mucilaginosa to antifungal agents: according to the review by Gomez-Lopez et al. [14], a good response to amphotericin B and flucytosine was observed, while susceptibility to azoles was poor. Interestingly, the susceptibility to voriconazole, which has rarely been tested against R. mucilaginosa isolates from clinical specimens [14], was also inadequate in our study. Our findings show that recovery of R. mucilaginosa from premature infants with clinical and laboratory signs of sepsis indicates a true infection for which specific antifungal treatment is required. Widespread antifungal prophylaxis, while providing protection against susceptible fungi, may contribute to the emergence of unusual, intrinsically resistant organisms.

Copyright information

© Springer-Verlag 2006