Nosocomial acquisition of Candida parapsilosis: An epidemiologic study

Nosocomial Acquisition of Cadida parapsilosis: An Epidemiologic Study VERONICASANCHEZ, M.D., Detroit, Michigan, JOSEA. VAZQUEZ,M.D., DANIELBARTH-JONES, M.P.H., LOUISEDEMBRY,M.D., Detroit, Michigan, andRoya/Dak, JACKD. SOBEL,M.D., Detroit, Michigan, MARCUSJ. ZERVOS,M.D.,

Michigan,

Detroit, Michigan, and Royal Oak, Michigan

OJ!MKTMQ The purpose of this study was to determine aspects of the epidemiology of nosocomial infection due to Candida parapailosds. Can&da specie43are important nosocomial pathogenq however, little epidemiologic information is available. PATIENTS AND =ODSz We prospectively culturtxlspecimensfrom98patientsadmittedtothe bonemarrowkanspbmtunitandamedkiueintensivecareunit(ICU)ofatertkycarehospital. &&mensfromhandsofpersonnelandenvironmsncal~W~also-~-tal cxlhalmWere~befOrepe~~Were&dmitted

toastudieduniLRe+Wk&m~analysis (RZA)ofchromawwlalDNAwasusedasatyp~ qWt4!!mtodeWmhietherela~of~ RESULTSC. parap&o&was idenMed from five patif3ut8$sixhandl?ultWsfromfour~staff, andtwoen -tall3u&ce%Allfivepatients hadIMgatlvehlitkllc&lllWl#UdaCquhX!dCjW ap?3&J&aft%radmis&mtothestudyunitaThere wereno&ni&xmtdiffer0ncesbelxveenpatien~ and~tE-olsUbje@Jinage,~diclegse,imm~therapy,andh&nunentation The duration of autibiotic therapy (mediam 328 velw311J3days#p=OS06)andthedurationinthe unit(me4~30.1ve~3us16.1days,p=0.048)was longer in patlenta than in controls. No common sourcewasiden~REAlwealedthreestrahl typee;-,Oll0f3tdltypewaSideolticalh

fourpatlentS#threestaffmemberi%andtwoenvironmentalCONCLUSION:Them redts suggest exogenous acquisition of C. pa.ra@osk Based upon isolation of identical patient straius of C. parap~&~& from inanimate surf&es before patients were admitted to a study unit, there is evidence that the organism may have been acquired from the hosFrom the Departments of Internal Medicine (VS. JAV. LD. JDS. MJZ) and Biostatistics @B-J). Division of Infectious Disease (VS, JAV. LD, JDS, MJZ). Wayne State University School of Medicine, Detroit, Michigan, and William Beaumont Hospital (JAV, DB-J, LD, MJZ), Royal Oak, Michigan. This study was supported in part by the William Beaumont Hospital Research Institute, Royal Oak, Michigan. Requests for reprints should be addressed to Marcus Zervos. M.D., Division of Infectious Diseases, William Beaumont Hospital, 3601 West 13 Mile Road, Royal Oak, Michigan 48072. Manuscript submitted February 28, 1992, and accepted in revised form November 6, 1992.

pital enviroument. The principal mechanism of . . trammmon was probably indirect contact via the hands of hospital personnel.

athogenic species are ubiquitous orP ganisms usually confined to human and animal reservoirs, but they can be recovered from soil, food, Candida

and the hospital environment. They are normal inhabitants of the female genital tract and the entire gastrointestinal tract, including the oropharynx, rectum, and perineum. Candida species are increasingly important nosocomial pathogens in immunocompromised patients, including intensive care and postoperative patients [l-23]. C. parapsilosis has been particularly associated with infections in hospitalized patients [12-231.Knowledge of the epidemiology of nosocomial Candida infection is incomplete because of the lack of a reliable typing system to evaluate strain relatedness. Previous typing systems have relied on serotyping and biotyping that do not reflect true strain differences. Recent advances in the use of restriction fragment analysis has facilitated the development of a typing system at a molecular level. These techniques have been used to study diverse groups of organisms including bacteria, viruses, and fungi [24-351.Molecular typing of Candida species (“DNA fingerprinting”) can differentiate closely related strains with phenotypic similarities [24-29,351.These tests have high discriminatory power. This typing system was used in this study to investigate aspects of the epidemiology of C. parapsilosis colonization and infection. In this prospective study, we used molecular typing of C. parapsilosis to further clarify the patient and hospital reservoirs of infection and modes of Candida transmission between high-risk patients. PATIENTS AND METHODS Harper Hospital is a 900-bed tertiary care teaching hospital in Detroit, Michigan. Patients in the medical intensive care unit (ICU) and bone marrow transplant (BMT) unit were studied. The medical ICU is an eight-bed private-room unit. Patients hospitalized in the ICU (ICUl) were transferred halfway through the study period to a separate IO-bed unit on a different floor with 4 open beds and 6 private beds. This new unit (ICU2) had previously been used as a June 1993

The Amerlcen Journal of Medicine

Volume 94

577

NOSOCOMIAL

CAND/DA

PARAPSILOSIS

/ SANCHEZ ET AL

surgical intensive care unit and underwent renovation prior to the transfer of patients. The same hospital personnel cared for patients in each ICU. The BMT unit is a 24-bed private room unit with a hospital staff different from that of the ICU. Patients were occasionally transferred between the BMT unit and the ICU, depending on the severity of their illness. Housestaff and medical students rotate between Harper Hospital and five other Wayne State University hospitals. We prospectively studied all patients who provided informed consent and who had been hospitalized in the BMT unit or ICU during a 7.5month period. Demographic and clinical data, including date of hospitalization, date of discharge, age, sex, ward, and medical service, were recorded by one of us (VS). The use of urinary catheterization and the dates and sites of vascular catheter placement were also noted. A careful record of bed assignments was kept. Movement within the hospital before discharge and exposure to antibiotic agents, cytotoxic agents, and immunosuppressive agents was also recorded. We defined colonization and infection according to established criteria: colonization was indicated when Candida species were isolated from a site without signs or symptoms of infection. Infection was diagnosed using criteria published for nosocomial infection: organism isolated from a body site concurrent with accompanying signs and symptoms of infection (e.g., fever greater than or equal to 38OC, leukocytosis in nonneutropenic patients, increased wound drainage, dysuria or pyuria, ophthahnitis). The severity of illness was determined using the criteria of McCabe and Jackson [36]. Death was considered related to Candida species if the patient died within 72 hours after a positive culture had been obtained on a sample from a normally sterile site, and the clinical profile was consistent with septicemia including hypotension, hypothermia or hyperthermia, tachycardia, tachypnea, and inadequate perfusion of at least one organ system. Specimens for culture were obtained from all patients and plated on Sabouraud dextrose agar. Clinical specimens were also plated on blood agar. Cultures were obtained when patients were admitted to the study unit, and then weekly for patients in the BMT unit and biweekly for patients in the ICU until discharge from the unit. Oropharyngeal, rectal, and urine cultures were obtained from all patients. Specimens from other sites such as wounds, vaginal areas, skin, and blood were also obtained when clinically indicated. The dates, sites, and colony counts of positive cultures for yeast were recorded. Oropharyngeal and rectal specimens were obtained using a sterile cotton swab. Urine samples were obtained through clean-catch voiding or were taken from syringe-aspirated catheters. An environmental culture survey performed in every patient room included at least one 578

June 1993

The American Journal of Medicine

Volume 94

of surfaces in contact with hands of personnel or patients, ventilator tubing, nebulizers, and humidifiers, one floor culture, and one sink faucet area culture. Specimens from the hands of hospital personnel were obtained by producing an impression with the hands on a Petri dish that contained Sabouraud dextrose agar. Cultures of selected environmental surfaces from the study units were done using Rodac plates (Falcon Inc, Oxnard, CA). Cultures of surfaces and hands were performed at 60-day intervals during the study period. Cultures of the surfaces in the medical ICU were obtained when the location of the unit was changed during the study period, including before any patients were hospitalized in the new unit. Patients from whom C. parapsilosis was isolated were matched by time of admission to the unit with three control patients without Candida species colonization. The statistical significance of data from patients and control subjects was assessed using tests of homogeneity of proportions and Fisher’s exact test for bivariate analysis of dichotomous outcome data. The significance of continuous variables was determined with the Mann-Whitney U test. A multivariate linear-regression model was used to analyze the independence of risk factors. All C. parapsilosis isolates were identified by negative germ tube formation, morphology, and Yeast API 20°C Systems (Sherwood Medical, Plainview, NY). Strains were initially isolated on Sabouraud dextrose agar (Difco, Detroit, MI), stored in a 1:l concentration of brain-heart infusion broth (Difco) and glycerol, and frozen at -7OOC until analysis. Control strains of C. parapsilosis were obtained from five patients located in a geographically separate hospital and C. parapsilosis ATCC strain #22019. Isolates were evaluated for molecular relatedness using restriction enzyme analysis (REA). REA was done by a blinded investigator (JV). For the isolation of DNA, yeast cells were grown on Sabouraud agar plates for 24 hours at 30°C. The method used to prepare the DNA is a modification of the method described by Petes et al [37]. The cells were suspended in 5.0 mL of 0.2M TRIS-HCl 5% mercaptoethanol and incubated in a water bath at 37OC for 30 minutes. After incubation, 10 mL of 1M sorbitol-0.004M potassium phosphate was added to the cells and pelleted. The cells were then resuspended in 2.5 mL of sorbitolpotassium phosphate-zymolyase 20T (ICN Biomedical, Lisle, IL) to promote spheroplast formation as per Torres-Bauxa and Riigsby [38]. The spheroplasta were then lysed using a solution of 0.2M sodium chloride-O.lM EDTA-5% sodium dodecyl sulfate-50mM TRIS (pH 8.5) -20 H pronase (Sigma, St. Louis, MO), and incubated at 60°C for 30 minutes in a water bath. Nucleic acids were extracted using a saturated solution of 1OmM TRIS-1mM EDTA (pH 8.3)-pheculture

NOBOCOMIAL CAiVNDAPARMSUOSS J BANCNEZETAL

nol-chloroform in a ratio of 25~2524and precipitated with two volumes of cold ethanol. The cell suspension was frozen for 1 hour at -7OOC. Afterwards, the nucleic acids were dried and resuspended in 1 mL of TE buffer (1mM TRIS, -O.lmM EDTA, pH 8.0) and reprecipitated with potassium acetate 3M (20 rL/0.5 mL) and two volumes of cold ethanol, then frozen for either 1 hour at -7OOC or overnight at -20°C. The nucleic acids were again dried in a vacuum oven for 30 minutes and resuspended in 106 rL of TE. The samples were stored at 4OC until electrophoresis. The concentration of DNA resulting from this procedure varies from 4 to 5 &pL. For the electrophoresis of samples, 5 & of DNA was digested with 50 U of the restriction endonucleasea EcoRI and MspI (BRL, Gaithersburg, MD), according to the recommendations of the manufacturer. Agarose gel electrophoresis of the purified DNA was then performed using a 0.7% agarose horizontal gel apparatus at 30V for 16 hours. The gel was stained in a solution of 0.5 &mL of ethidium bromide for 20 minutes, then destained in water for 60 minutes before being photographed. Grouping was done by visual inspection of the banding patterns.

RESULTS During the study period, we followed 98 patients with 106 admissions (49 males and 49 females) hospitalized in the BMT unit and the ICU. Six patients had two admissions each, and one patient had a total of three admissions. Patients ranged in age from 5 to 66 years (mean: 46.8 years). Six patients were transferred between the BMT unit and the ICU during their hospital stay. Patients in the ICU were hospitalized a mean of 11.3 days (range: 2 to 43 days) and in the BMT unit for a mean of 27.4 days (range: 3 to 73 days). Five patients were found to have C. parapdosis at some time during their hospital stay. The recovery of C. parapsilosis by body site and as a proportion of total Candida isolatesis shown in Table I. We recovered a total of eight isolates of C. parapsilosis from the five different patients. The organism was most frequently isolated from the oropharynx and stool. No patient had positive blood cultures. All five patients had negative initial cultures and grew colonies of C. parapsilosis only after admiiion to the study unit (Table II). Two patients were located in the BMT unit, and three patients were located in ICU2 (Figure 1). Initial sites of colonization in patients who acquired the C. parapsiZosis colonies while the in hospital were as follows: the rectum in two patients, the oropharynx in two patients (one patient also had vaginal colonization), and a sternal wound in one patient. There were no clinical infections or deaths related to C. parapsilosis. Three patients had C. parapsiZosis isolated just prior to discharge, and no repeat

TABLE I Recoveryof C. pampsisilosis by Body Site and as a Percentaged Total study Candida SpKies koMes

Orophatynx

3

119

2.5

Stool

3

138

2.2

Urine

0

77

0

Vaginal

1

36

2.8

Blood

0

4

Wound

1

14

Sputum

0

2

Total

a

390

0 7.1

0 2.0

TABLE II Patient LocationWith Acquisitionof C. pafapsibsis

Unit

TotalPatients at Unit

WiiNew Acquisitien

ICU2 BMT

44 54

;

:

Total

98

5

0

With Cobniis at Transferto Unit

U2 = intensivecare unit 2; BMT = bane marrow transplant.

cultures were done. The other two patients had repeat cultures from which C. parapsilosis wannot isolated. One patient received systemic amphotericin B for 6 days prior to the first culture of C. parapsilosis. None of the other patients received topical or systemic antifungal therapy prior to culturing of C. parapsilosis. Patient and control demographic and risk factors are summa&ed in Table III. The mean duration of hospit&ation, the duration in the study unit, and the duration of antibiotic therapy prior to the isolation of C. parapsilosis were significantly longer for patients than for control subjects. Antibiotic duration and duration in the study unit were not independently associated with C. parapsilosis. There were no differences between patients and controls with regard to prior antibiotic type admmistered, severity of illness, antifungal use or duration, immunosuppressive use or duration, catheter use or duration, underlying illness, or the presence, severity, or duration of granulocytopenia. Six hand cultures from four hoepital staff were positive for C. parapsilosis. Colonies of C. parapsilosis were found on two environmental surfaces cultured in ICUS. The environmental surfaces that were found to contain C. parapsilosis were those that were frequently touched by hospital personnel or in open conJune 1993

The Americen Journal of Medlclne

Volume 94

579

NOSOCOMIAL

CANDIDA

PARAPSILOSIS

/ SANCHEZ ET AL

TABLE III

BMTUJ

Patient and ControlDemographicand Risk Factor Characteristics

14

Patients

2

(n = 5)

Controls (n = 15)

V&e

16

ZZ”

Range

Sex (F/M) Duration in unit(d) Mean Range Prior antibiotics 2l

Duration antibiotics (d) Median Range

22 23 24

47.2 25-88

4/l

817

30.1 11-51

16.1 3-41

5

p = 0.048

13

32.8 11-45

11.8 o-61

Prior antiiungal therapy (topical or systemic)

1

7

Immunosuppressivetherapy

2

6

Granulocytopania

2

4

:

2”

2 -

7 2

Underlying illness Cardiac disease Neurologicdisease Hematologic Malignancy Other

Figure 1. Schematic floor plan of the medical intensive care unit (ICU2) and bone marrow transplant (EMT) unit. H = patient with acquisition of C. parapsilosis in the unit; * = positive environmental culture for C. parapsilosis; ** = hand culture positive for C. parapsilosis. Numbers 1 to 24 and 25 to 34 denote patient room numbers.

38.4 1965

p = 0.05

strains, were separate strain types having unique restriction fragment patterns. Evaluation of isolates colonizing different body sites within an individual patient showed that this patient was colonized with two strain types of C. parapsilosis in the oropharynx. One of the pharyngeal strain types was the same as that cultured from the vaginal specimen of the same patient.

COMMENTS tact with patients. Colonies of C. parapsilosis were C. parapsilosis is an infrequent opportunistic cultured from a sink and from an air-conditioning pathogen but is of special interest because of its ability to infect seriously ill hospitalized patients. C. parvent in the ICU. These isolates were obtained before patients were moved into the new ICU. A patient apsilosis does not appear to be a commensal organhospitalized in the room where C. parapsilosis was ism. Isolates of C. parapsilosis have been recovered cultured from the air-conditioning vent had negative from urine, oropharynx, and stool of hospitalized painitial cultures but later had an identical strain of C. tients [l-21]. Dick et al [S] found C. parapsilosis to parapsilosis first cultured from a sternal wound. be an infrequent isolate representing 4% of yeast isoThere was no common source (vehicle) identified. lates recovered from oncology patients at the Johns All C. parapsilosis isolates were evaluated by Hopkins Hospital. Bloodstream infection due to C. REA. There were eight isolates from five patients, six parapsilosis is not uncommon and has been identihand cultures from four hospital staff, two environfied in up to 28% of all fungemias [13]. C. parapsilosis mental isolates, six control strains, and one ATCC appears to be tiquely related to intravascular cathestrain. REA revealed a total of 3 different restriction ter infection. Extravascular infection and invasive enzyme patterns (strain types) among the 23 strains candidiasis are uncommon. examined. One strain type was identical in four paSerious infections due to C. parapsilosis have been tients (three in ICU2), three staff (two nurses and one described among intravenous drug users, recipients of respiratory therapist), and two environmental sur- parental nutrition solutions, persons with intravascufaces (sink and air vent in ICUB) (Figure 2). The lar blood pressure monitoring transducer devices, and isolates from the remaining patient, staff, and control hemodialysis recipients, and are a cause of postopera580

June 1993

The Amerfcan

Journal

of Medlclne

Volume

94

NOSOCOMML CANDIDA PARAPSJLOSJS/ SANCHEZ ET AL tive

infections, including endophthahnitis [12-231. Other patients who are at risk for C. parupsilosis colonization and infection are immunocompromised hospitalized patients who are usually receiving prolonged therapy with broad-spectrum antibiotics. In the current study, case-control analysis found that patients with C. parupsilosis had a longer duration of hospital&ion and a longer duration of antibiotic therapy when compared with patients from whom Candidu was not cultured. These findings are similar to earlier reports of infection with Candida albicans [l-12]. In this study, there were no differences between patients and control subjects with regard to antifungal prophylaxis, immunosuppressive or cytotoxic use, catheter use, hyperalimentation, other instrumentation, or severity of underlying ilhress. Little is known about reservoirs of C. parapsilosis, which most likely include a complex interaction of environmental and human sources. Meunier-Carpentier et al [12]found that of 13 patients with C. parapsilosis fungeti, none had the organism isolated from other sit.ee before the onset of fungemia. This differs from bloodstream infection due to C. albicans or Candida tropicalis in which 62% and 3% of patients, respectively, with identical conditions were found to have positive cultures from other sites prior to the onset of fungemia [12].These data suggest that the organism causes fungemia directly through access from an environmental or other exogenous source. The unique role for the environment as a potential reservoir for Candida species is further supported by fmdings in the current study showing culture of C. parupsilosis from an air-conditioning vent before patients were hospitalized in a new ICU. Subsequent cultures from a sternal wound of a patient hospitalised in that room grew an identical strain type of C. parupsilosis, which suggest potential acquisition of the organism from the environmental surface. This observation is significant because organisms isolated from the inanimate hospital environment are generally believed to contribute negligibly to endemic nosocomial infection [38].This conclusion is supported by data showing that despite differences in flora isolated from environmental surfaces in different hospitals, the incidence and type of organisms causing nosocomial infection remain the same. Furthermore, although infecting strains can be cultured from environmental surfaces, it is likely that the environment becomes passively contaminated by organisms from patients (%mocent bystanders”). In this study, culture of surfaces in the unit prior to the admission of patients afforded a unique opportunity to study the role of the environment. The results of this study suggest that Candida may be unlike other nosocomial pathogens and can possibly be acquired directly or indirectly from contaminated environmental surfaces.

Flgure 2. Agarose gel electrophoresis and restriction enzyme analysis of purified DNA from strains of C. parapsilosis from five patients showing three strain types. Lanes B through F were digested with Mspl, and lanes l-l through L were digested with EcoRI. Lanes A and G show bacteriophage X for size reference. Size markers are shown to the left.

There have been occasional reports of nosocomial clusters of infection due to Candida species with suspetted patient-to-patient spread or suspected acquisition from environmental sources. Y&velvecases of C. parapsilosis fungemia in association with an infected blood pressure transducer in an ICU [23]and 15 cases of C. parapsilosis endophthalmitis secondary to an infected irrigating solution [17]suggest exogenous acquisition. Other studies have implicated carriage on the hands of hospital personnel as possible sources of outbreaks. Eight cases of C. tropicalis mediastinitis following thoracic surgery were traced to a single operating room nurse who had isolation of C. tropicalis from both the oropharynx and hands [40], and an outbreak of C. albicans in an ICU [41] suggested exogenous acquisition. This study demonstrates exogenous acquisition of C. parapsilosis within the hospital environment and traces the route of acquisition of this pathogen. The isolation of identical strains of C. parapsilosis from staff and patients suggests that the major route of trensmission is indirect contact between patients. Transient carriage on the hands of hospital personnel was suggested, but we cannot exclude the possibility that the hospital staff may have acted ss “innocent bystanders” as opposed to vectors. Accordingly, the epidemiology of nosocomial C. parapsilosis infection resembles that of methicillin-resistant staphylococci, multi-drug-resistant enterococci, and gram-negative bacilli [30-331.The most important route of spread of these resistant bacteria in the hospital is from an infected patient by transient carriage on the hands of personnel. The role of personal carriage of Candida in the dissemination of strains remains to be clarified. Moreover, Candida organisms were readily transmitted between patients in adjacent beds within the June 1993

The Amerlun

Journal of Medlclne

Volume 94

581

NOSOCOMIAL

CANDIDA

PARAPSllDSIS/

SANCHEZ

ET AL

ICUs. Proximity to a patient with infection or colonization increased the risk of acquisition of Candida. The source (reservoir) of C. parapsilosis is probably the respiratory tract, gastrointestinal tract, or genitourinary tract of colonized patients. Although no de% nite symptomatic infections occurred in the patients who became colonized in this study, the ability of C. parapsilosis to cause serious infection in hospitalized patients has been amply demonstrated in other studies [B-23]. The epidemiology of nosocomial C. parapsilosis infection is complex. Potential hospital reservoirs -include colonized patients, hospital personnel, and the inanimate hospital environment. The mechanism of acquisition of C. parapsilosis in our patients remains uncertain; however, indirect contact transmission is most likely. A potential role for the environment as a source for Candida cannot be excluded. Our findings demonstrate that nosocomial C. parapsilosis colonization is a result of the exogenous acquisition of strains. The potential for dissemination of C. parapsilosis in hospitalized patients is disturbing. Further prospective studies are needed to more accurately define patient and hospital reservoirs of infection, as well as the mode of transfer and measures for preventing spread.

in a patient receiving chronic hemodialysis. J Rheumatol

1988; 15: 72-4.

16. McCray E. Rampell N. Solomon SL, Bond WN, Martone WJ, D’Day D. Outbreak of Candida parapsihsis lens implantation.

endophthalmitis

after cataract extraction and intraocular

J Clin Microbial 1986; 24: 625-8.

17. Stern WH, Tamura E, Jacobs RA, eta/. Epidemic postsurgiwl Candidaparapsk losis endophthalmitis. Ophthalmology

Clinical findings and management

of 15 consecutive wses.

1985; 92: 1701-g.

18. Law EJ, Holder IA, MacMillian BG. Candida parapsilosis

fungemia in burn

patients: report of three cases. Burns, Including Thermal Injury 1984; 10: X3-6. 19. Solomon SL. Khabbaz RF, Parker RH, Anderson RL, Geraghty MA. An outbreak of Candida parapsiloss

bloodstream

infections in patients receiving parenteral

nutrition. J Infect Dis 1984; 149 98-102. 20. Younkin S, Evarts CM, SteigMgel RT. Candida parapsilosis infection of a total hip-joint replacement: successful reimplantation B and 5fluorocytosine.

after treatment with amphoteridn

A case report. J Bone Joint Surg [Am] 1984; 66: 142-3.

21. Faix FIG. Candida parapsilosis meningitis in a premature infant. Pediatr Infect Dis 1983; 2: 462-4 22. Walter B. Nolte J, Kappe R, Muller J. Relapsing fur-gal septicemia caused by Candidaparapsilosisfollowing

staphylcwccal

septicemia. Mycoses 1988; Suppl 1:

54-B. 23. Weems JJ, Chamberland ME, Ward J. Wtlly M, Padnye AA, Solomon SL. Candi da parapsilosis fungemia associated with parenteral nutrition and contaminated blood pressure transducers.

J Clin Microbial 1987; 25: 1029-32.

24. Kaufmann CS, Merz WG. Electrophoretic Clin Microbiil

karyotypes of Torulopsisglabrata.

J

1989; 27: 2165-8.

25. Matthews R, Bumie J. Assessment of DNA fingerprinting for rapid identiiwtion of outbreaks of systemic wndidiasis.

BMJ 1989; 298: 354-7.

26. Merz WG, Conneliy C, Hieter P. Variation of electrophoretic

karyotypes among

clinical isolates of Candida albicans. J Clin Microbial 1988; 26: 842-5. 27. Stevens DA, Odds FC. Scherer S. Applications of DNA typing methods Candida albicans epidemiology and correlations With phenotype.

to

Rev Infect Dis

1990: 12: 25866. 28. Su CS, Meyer SA. Restriction endonudease

REFERENCES 1. BodeyGP. Candkiiiis

analysis of mitochondrial DNA from

Candida parapsilosis and other Candida species. Yeast 1989; 5: S355-60. of a typical Candida tropicais and

29. Schlitzer RL, Ahearn DG. Characterization in cancer patients. Am J Med 1984: 77(Suppl4D):

other uncommon

13-9.

clinical yeast isolates. J Clin Microbial 1982; 15: 51 l-6.

3O.Zervos

MJ, Dembinski S, Mikesell T, Schaberg DR. High-level resistance to

frequency, earlier onset, and resuits of therapy. Rev Infect Dis 1985; 7: 646-55.

gentamicin

in Streptococcus

3. Gains JO. Remington JS. Disseminated candid&is

acquisition of infection. J Infect Dis 1986; 153: 1075-82.

2. Horn R. Wong B. Kiehn TE, Armstrong D. Fungemia in a wnwr

hospital: changing

in the surgical patient. Sur-

risk factors and evkienw

31. Farrar WE Jr. Molecular analysts of plasmids in epkiemiologiwl

gery 1972; 72: 730-6. 4. Klein JJ, Watanakunakorn clinical signiinw.

faecak

C. Hospital acquired fungemia. Its natural course and

Am J Med 1979; 67: 51-8.

for exogenous investigation. J

Infect Dis 1983: 148: l-6. 32. Schaberg DR. Tompkins LS, Falkow S. Use of agarose gel electrophoresis

5. Sobet JD. Vazquez J. Candidemia and systemic wndkfiasis. Semin Respir Infect

plasmid deoxyribonucleic acid to finger print gramnegative

of

bacilli. J Clin Miaobiol

1990; 5: 123-37.

1981;13: 1105-8.

6. Dick JD. Merz WG, Sara1 R. Incidence of polyene-resistant yeasts recovered from

33. McGowan JE Jr, Terry PM, Huang T Sr. Houk CL, Davies J. Nosocomial infec-

diniwl

specimens. Antimicrob Agents Chemother

tions with gentamicin

1980; 158-63.

resistant Staphylococcus

aureus: plasmid analysis as an

7. Bross J, Talbot GM, Maislin G, Hurwitz S, Strom BL. Risk factors for nosocomial

epidemiologic tool. J Infect Dis 1979: 140 86472.

wndiiemia:

34. Vazquez JA, Bed&y

a case-control study in adults without leukemia. Am J Med 87: 614-20.

8. Burnie JP. Matthews R, Lee W, Philpott-Howard nosocomial systemic wndidiasis.

Epiwmiol

J, Brown R. Four outbreaks of

A, Sot& JD, Zervos MJ. Comparison of restriction enzyme

analysts versus pulsed-field gradient gel electrophoresis

as a typing system for

Candida albicans. J Clin Microbial 1991;291962-7.

Infect 1987; 99: 201-11.

9. Vaudry WL, Tiemey Al. Wenman WM. Investigation of duster of systemic Candk

35. Fox BC, Mobley HLT, Wade JC. The use of a DNA probe for epidamiologiwl

da ahiwns

studies of wndidiasis

infections in a neonatal intensive care unit. J Infect Dis 1988; 158:

1375-9.

in immunocompromised

hosts. J Infect Dis 1989; 159

488-94.

10.Wey SB, Motomi M, Pfaller MA, Woolson RF, Wenzel RP. Hospital-acquired

36. McCabe WR. Jackson GG. Gram-negative

wndkfemia:

Arch Intern Med 1962:110:847-55.

the attributable mortality and excess length of stay. Arch Intern Med

1988; 148: 2642-7. 11. Wey SB. Motomi hospital-acquired

M, Pfaller MA. Woolson RF. Wenzel RP. Risk factors for

wndkfemia:

a matched

wsecontrol

study. Arch Intern Mad

1989; 149 2349-53. 12. Meunier-Carpantier

bacteremia.

I. Etiology and ecology.

37. Petes TD, Hereford LM, Skryabin K. Characterization

of two types of yeast

ribosomal DNA genes. J Bacterial 1978; 134: 295-x)5. 38. Torres-Bauza Candida albiwns.

F, Kiehn TE, Armstrong D. Fungemia in the compromised

U, Riggsby WS. Protoplasts from yeast and mycelial forms of J Gen Microbial 1980;119:341-9.

39. Maki DG, Alvarado CJ, Hassemr CA, Zi!z MA. Relation of the inanimate hospital

host. Am J Med 1981;71: 353-70.

environment

13. Mayer AR, Brown A, Weintraub RA. Ragni M, Postic B. Successful medical

40. lsenberg HD, Tucoi V. Cintron F, Singer C, Weinstein GS, Tyras DH. Singie

therapy for endowrditis

due to Candida parapsiosis.

to endemic nosocomial infection, N Engi J Med 1982: 307: 15626.

source outbreak of Candida fropicaliscompliwting

Chest 1978:73: 546-9.

coronary bypass surgery. J Clin

14. D’Day DM, Head WS, Robinson RD. An outbreak of Candida parapsilosis en-

Microbial 1989; 27: 2426-B.

dophthalmitis; analysis of strains by enzyme profile and antifungal suswptibilii.

41. Bumie JP, Odds FC. Lee W, Webster C, Williams JD. Outbreak of systemic

J Ophthalmol

Candida albicans in an intensive care unit caused by cross infection. BMJ 1985;

15. DeClerckL. Dequeker J, Westhovens R, HaJustaine

582

Br

1987; 71: 126-9.

June 1993

The American

Journal

D. Candida parapsiosis

of Medicine

Volume

290 746-8.

94