National Epidemiology of Mycoses Survey: A Multicenter Study of Strain Variation and Antifungal Susceptibility Among Isolates of Candida Species

MYCOLOGY

National Epidemiology of Mycoses Survey: A Multicenter Study of Strain Variation and Antifungal Susceptibility Among Isolates of Candida Species M. A. Pfaller, S. A. Messer, A. Houston, M. S. Rangel-Frausto, T. Wiblin, H. M. Blumberg, J. E. Edwards, W. Jarvis, M. A. Martin, H. C. Neu, L. Saiman, J. E. Patterson, J. C. Dibb, C. M. Roldan, M. G. Rinaldi, and R. P. Wenzel The National Epidemiology of Mycoses Survey (NEMIS) involves six academic centers studying fungal infections in surgical and neonatal intensive care unit (ICU) patients. We studied variation in species and strain distribution and antifungal susceptibility of 408 isolates of Candida spp. Candida spp. were isolated from blood, other normally sterile site cultures, abscesses, wounds, catheters, and tissue biopsies of 141 patients hospitalized in the surgical (107 patients) and neonatal (34 patients) ICUs of medical centers located in Oregon, Iowa, California, Texas, Georgia, and New York. Isolates were also obtained from selected colonized patients (16 patients) and the hands of health care workers (27 individuals). DNA typing was performed using pulsed field gel electrophoresis, and antifungal susceptibility to amphotericin B, 5-fluorocytosine, fluconazole, and itraconazole was determined using National Committee for Clinical Laboratory Standards (NCCLS) meth-

ods. Important variation in susceptibility to itraconazole and fluconazole was noted: MICs of itraconazole ranged from 0.25 mg/mL (MIC90) in Texas to 2.0 mg/mL (MIC90) in New York. Similarly, the MIC90 for fluconazole was higher for isolates from New York (64 mg/mL) compared to the other sites (8–16 mg/mL). In general, DNA typing revealed patient-unique strains; however, there were 13 instances of possible crossinfection noted in 5 of the medical centers. Notably, 9 of the 13 clusters involved species of Candida other than C. albicans. Potential transmission from patient-to-patient (C. albicans, C. glabrata, C. tropicalis, C. parapsilosis) and health care worker-to-patient (C. albicans, C. parapsilosis, C. krusei) was noted in both surgical ICU and neonatal ICU settings. These data provide further insight into the epidemiology of nosocomial candidiasis in the ICU setting. © 1998 Elsevier Science Inc.

From the University of Iowa College of Medicine, Iowa City, Iowa (MAP, SAM, AH, MSR-F, TW); Emory University School of Medicine, Atlanta, Georgia (HMB, WJ); University of California School of Medicine, Los Angeles, California (JEE); Centers for Disease Control and Prevention, Atlanta, Georgia (WJ); Oregon Health Sciences University, Portland, Oregon (MAM); Columbia University, New York, New York (HCN, LS); University of Texas Health Science Center, San Antonio, Texas (JEP, JCD, CMR, MGR); Medical College of Virginia, Virginia Commonwealth University, Richmond, Virginia (RPW), USA. Address reprint requests to Dr. M. A. Pfaller, University of Iowa College of Medicine, Department of Pathology, Room C606 GH, 200 Hawkins Drive, Iowa City, IA 52242, USA. Received 11 November 1997; revised and accepted 12 December 1997.

INTRODUCTION

DIAGN MICROBIOL INFECT DIS 1998;31:289 –296 © 1998 Elsevier Science Inc. All rights reserved. 655 Avenue of the Americas, New York, NY 10010

The incidence of nosocomial fungal infections has increased dramatically in the USA over the past two decades, with a consequent rise in related mortality and prolonged hospitalizations (Beck-Sague and Jarvis 1993; Emori and Gaynes 1993; Fisher-Hoch and Hutwagner 1995; Pfaller 1995a; Pinner et al. 1996; Pittet and Wenzel 1995; Wey et al. 1988). Numerous factors have contributed to this trend: 1) most notably, a growing population of immunocom-

0732-8893/98/$19.00 PII S0732-8893(97)00245-9

M.A. Pfaller et al.

290 promised or immunosuppressed patients whose mechanisms of host defense have been impaired by primary disease states (e.g., AIDS, cancer, or diabetes); 2) the use of broad-spectrum antibiotics and immunosuppressive therapeutic regimens (e.g., cytotoxic chemotherapies); and 3) the use of new and aggressive therapeutic strategies or life support systems (Beck-Sague and Jarvis 1993; Fridkin and Jarvis 1996; Jarvis 1995; Pfaller 1995a; Wenzel 1995; Wey et al. 1989). Particularly prominent are the sharp increase in the incidence of candidal blood stream infections among hospitalized patients and the rising importance of species of Candida other than C. albicans (Banerjee et al. 1991; Fisher-Hoch and Hutwagner 1995; Fridkin and Jarvis 1996; Nguyen et al. 1996; Pfaller 1996). Although the understanding of the epidemiology of fungal infections has grown appreciably in recent years, rapid changes in this field mandate close surveillance to determine actual rates of infection and to identify emerging pathogens and potential risk factors (Fridkin and Jarvis 1996; Jarvis 1995; Wenzel 1995; Pfaller 1995a). Two populations that appear to be at great risk and requiring close monitoring are patients in surgical and neonatal intensive care units (ICUs) (Fisher-Hoch and Hutwagner 1995; Fridkin and Jarvis 1996). These patients account for almost 50% of nosocomial fungal infections nationwide (Beck-Sague and Jarvis 1993; Jarvis 1995; Pfaller 1995a; Wenzel 1995). There is a continuing need to monitor the incidence of such infections, the spectrum of fungal pathogens involved, and the risk factors for acquiring fungal infections in these populations. These issues are best addressed within the context of a collaborative, multicenter epidemiological study using prospective surveillance techniques as well as standardized protocols and definitions of infection. The National Epidemiology of Mycoses Survey (NEMIS) was instituted in 1993 for the detailed assessment of fungal infections among patients hospitalized in surgical and neonatal ICUs (Pfaller 1995b). Among the objectives of the NEMIS study was the need to define the spectrum of pathogens causing nosocomial fungal infections in these critically ill populations, to characterize the organisms with respect to their susceptibility to commonly used antifungal agents, and to define the molecular epidemiology of these infections with respect to endogenous sources and cross-infection. In the present study, we focused on infections caused by Candida spp. and described the variations in species and antifungal susceptibility among isolates from the six study sites. Molecular typing methods were used to identify individual strains and to detect possible crossinfection.

METHODS Study Sites The six geographically distinct study sites represented the states of Georgia (GA; Grady Memorial Hospital, Emory University School of Medicine, Atlanta, GA), California (CA; University of California School of Medicine, Los Angeles, CA), Oregon (OR; Oregon Health Sciences University, Portland, OR), New York (NY; Columbia Presbyterian Medical Center, New York, NY), Texas (TX; University of Texas Health Science Center, San Antonio, TX), and Iowa (IA; University of Iowa College of Medicine, Iowa City, IA).

Surveillance Prospective surveillance was conducted over a 2-year period (1993–1995) on all patients hospitalized at least 48 h in the surgical and neonatal ICUs (SICU and NICU, respectively) of each study site. Nosocomial infections were identified using definitions recommended by the Centers for Disease Control (CDC) (Garner et al. 1988). Microbiologic studies during the course of this survey included weekly surveillance cultures of stool and urine for Candida spp. (rectal swab only in NICU), as described previously (Pfaller et al. 1987). Additional cultures of the hands of health care workers (HCW) were obtained using the brothbag method (Strausbaugh et al. 1996). Hand cultures were obtained on a monthly basis and whenever an episode of candidemia was recognized. These isolates plus all Candida spp. from infected patients (blood, normally sterile body fluids, abscesses, tissue biopsies, catheters, wounds) were saved and sent to the laboratory at Iowa for banking and further analysis.

Organisms A total of 408 isolates of Candida spp. were selected for analysis. The selected isolates represent those obtained from 141 patients (107 SICU and 34 NICU) with nosocomial candidiasis (145 infections, 365 isolates), 16 colonized patients (16 isolates), and 27 HCWs (27 isolates from hand cultures). These isolates included 261 from the SICU and 147 from the NICU. In selecting these isolates for the present study, we chose the etiologic agent of each infection (at least one isolate per infected patient). Colonizing as well as infecting isolates from patients were included to demonstrate the relationship between colonizing and infecting strains. The isolates from colonized uninfected patients in the same ICUs and the hands of HCWs were included because they were related to specific infections. All isolates were iden-

National Epidemiology of Mycoses Survey

291

TABLE 1 Sites of Nosocomial Candidiasis Observed in the Surgical and Neonatal ICUs of 6 Academic Medical Centers No. of infections by Location (n) Site of Infection

OR (22)

CA (34)

IA (25)

TX (21)

GA (31)

NY (12)

All (145)a

Blood Sterile body fluids Abscess Tissue Wound Catheter

13 2 6

9 8

15 6

11 6 1

28 3

8

84 25 8 4 23 1

a

1

4 1

17

3

2 1

Represents 145 infections in 141 patients.

TABLE 2 Species Distribution of Nosocomial Isolates of Candida spp. from 6 Academic Medical Centers % by Location Species C. albicans

C. glabrata

C. tropicalis

C. parapsilosis

C. krusei

Candida spp.

Total isolates

ICU

OR

CA

IA

TX

GA

NY

All

SICU NICU All SICU NICU All SICU NICU All SICU NICU All SICU NICU All SICU NICU All

54 26 41 8 0 4 14 0 8 8 73 38 10 0 5 6 1 4

60 67 61 8 0 7 20 0 17 6 33 10 6 0 5 0 0 0

78 100 81 9 0 7 7 0 6 4 0 4 0 0 0 4 0 0

64 40 60 28 0 23 4 0 4 4 60 13 0 0 0 0 0 0

46 56 50 14 11 13 18 0 11 11 33 20 0 0 0 11 0 6

57 97 73 41 0 24 0 0 0 2 3 3 0 0 0 0 0 0

59 51 56 16 1 11 11 0 7 6 47 20 4 0 3 4 1 3

SICU NICU All

91 78 169

35 6 41

45 9 54

25 5 30

21 18 39

44 31 75

261 147 408

tified to species by conventional methods (Warren and Hazen 1995) and were stored as suspensions in sterile water at ambient temperature until used.

Antifungal Susceptibility Testing The in vitro susceptibility of the 408 isolates to a panel of four antifungal agents (amphotericin B, 5-fluorocytosine [5FC], fluconazole, and itraconazole) was determined using a standardized broth microdilution method (NCCLS 1997). Quality control (QC) was ensured by testing the QC strains recommended by the National Committee for Clinical Laboratory Standards (NCCLS), Candida parapsilosis ATCC 22019 and C. krusei ATCC 6258 (NCCLS 1997; Pfaller et al. 1995).

Molecular Typing Molecular typing of 396 (97.1%) of the isolates was accomplished by using restriction endonuclease analysis (REA) followed by pulsed field gel electrophoresis (PFGE), electrophoretic karyotyping (EK), and REA followed by conventional electrophoresis (C. krusei only).

REA with PFGE Isolates were typed by REA with PFGE as described previously (Hollis et al. 1995). Briefly, agarose plugs containing DNA were placed in 100 mL of buffer containing 20 U of a low-frequency cleavage restriction endonuclease (BsshII or SfiI); overnight digestion

M.A. Pfaller et al.

292

TABLE 3 NEMIS Study: Institutional Differences in In Vitro Susceptibility of Nosocomial Isolates of Candida spp. to 4 Antifungal Agents MIC (mg/mL) Study Site

N

OR

164

CA

39

IA

54

TX

30

GA

39

NY

75

Range

50

90

0.25–2.0 0.06–256 0.12–128 0.015–16 0.5–1.0 0.06–256 0.12–256 0.015–16 0.25–1.0 0.12–.256 0.12–256 0.015–16 0.5–1.0 0.06–256 0.12–256 0.015–16 0.5–2.0 0.25–8.0 0.12–128 0.015–16 0.5–1.0 0.06–256 0.12–256 0.007–16

1.0 0.25 0.5 0.06 1.0 0.12 0.25 0.015 1.0 0.25 0.5 0.06 1.0 0.12 0.25 0.015 1.0 0.25 0.5 0.12 1.0 0.5 0.25 0.03

1.0 2.0 8.0 0.5 1.0 2.0 8.0 0.5 1.0 0.25 16 1.0 1.0 1.0 8.0 0.25 1.0 2.0 8.0 1.0 1.0 2.0 64 2.0

Antifungal Agent Amphotericin 5FC Fluconazole Itraconazole Amphotericin 5FC Fluconazole Itraconazole Amphotericin 5FC Fluconazole Itraconazole Amphotericin 5FC Fluconazole Itraconazole Amphotericin 5FC Fluconazole Itraconazole Amphotericin 5FC Fluconazole Itraconazole

B

B

B

B

B

B

was performed as recommended by the manufacturer. PFGE was performed using a contour clamped homogeneous electric field (CHEF) electrophoresis system (CHEF-DRII; BioRad, Richmond, CA) at 200 V and 13°C for 24 h in 1% agarose gel, with switch times ramped from 10 to 90 s. Gels were stained with ethidium bromide and photographed under ultraviolet light.

REA with Conventional Electrophoresis Isolates of C. krusei were typed by REA using the restriction endonuclease HinfI followed by conventional electrophoresis (1% agarose, 40 V, 26 h) as described by Berrouane et al. (1996).

Analysis Analysis of REA and EK profiles was performed by visual inspection of photographs of ethidium bromide-stained gels to detect similarities and differences. Isolates were considered different if banding patterns differed by more than one (EK) or two (REA) readily detectable bands. To achieve maximum strain discrimination, the results of REA and EK analysis were combined to achieve a composite DNA type. Comparisons of EK and REAG patterns were limited to isolates from patients in the same institution. No attempt was made to compare DNA types from patients at different institutions.

RESULTS AND DISCUSSION Electrophoretic Karyotyping (EK) Analysis Isolates were typed by EK analysis as described by Pfaller et al. (1994). Briefly, agarose plugs containing DNA were inserted into an agarose gel (1%) and the intact chromosomes were resolved by PFGE using the CHEF-DRII system. PFGE was conducted at 150 V and 13°C for 36 h. Gels were stained with ethidium bromide and photographed under ultraviolet light.

A total of 145 nosocomial infections due to Candida spp. in 141 patients was observed over the 2-year study period (Table 1). Among these infections, 111 occurred in 107 patients hospitalized in the SICU and 34 occurred in 34 patients in the NICU. Blood stream infection was most common and accounted for 58% of all nosocomial candidal infections (49% SICU and 88% NICU) followed by infection of normally sterile body fluids (17%) and wound infections (16%).

National Epidemiology of Mycoses Survey The distribution of the different species of Candida isolated in the six study sites is shown in Table 2. Of the 408 isolates, 261 (64%) were from the SICU and 147 (36%) were from the NICU. Although the majority of the isolates were C. albicans (56%; 51% NICU and 59% SICU), the proportion varied from a low of 41% in OR to a high of 81% in IA. The differences in the frequency of C. albicans were greatest in the NICU setting where only 26% of isolates in OR were C. albicans, compared with 97% in NY and 100% in IA. Among the non-albicans species, C. parapsilosis was most common in OR and GA, whereas C. glabrata predominated in TX and NY. C. tropicalis was most common in CA. Notably, C. parapsilosis accounted for over 30% of the isolates in four of the six NICUs. Other species of Candida were remarkably absent from the NICUs. C. glabrata was only observed in the GA NICU, whereas C. tropicalis and C. krusei were not detected in any of the NICUs. In the SICUs, C. glabrata was most common in TX (28%) and NY (41%), whereas C. tropicalis accounted for 14%, 18%, and 20% of the isolates in OR, GA, and CA, respectively. Although C. krusei was only isolated in two of the six study sites, it accounted for 10% of the isolates from the SICU in OR. The in vitro susceptibility of the 408 isolates of Candida spp. to 4 commonly used antifungal agents is shown in Table 3. Only minor variations in the MICs of amphotericin B and 5FC were observed among the six study sites. The MIC90 for 5FC was four- to eightfold lower for isolates from IA compared to the other study sites. It is notable that 5FC is used rarely for treatment of candidiasis (in combination with amphotericin B) at the University of Iowa. Variation in MICs for the azoles was observed among the six study sites. The MIC90 for itraconazole ranged from a low of 0.25 mg/mL in TX to 2.0 mg/mL for isolates from NY. Similarly the MIC90 for fluconazole was significantly higher for isolates from NY (64 mg/mL) compared to the other sites (8 –16 mg/mL). The elevated azole MICs among NY isolates were accounted for by several isolates of C. glabrata (SICU) and by isolates of C. albicans obtained from a patient on azole (fluconazole) therapy (NICU). A great deal of strain diversity was detected by DNA typing methods (Table 4 and Figures 1 and 2). In general, each patient was infected by his or her own distinct DNA type of Candida (Figure 1). In cases in which multiple cultures were positive, the same DNA type was isolated repeatedly over time (Figures 1 and 2). However, several patients had more than one strain present at different times and one patient sustained three different episodes of fungemia with three different species of Candida.

293 TABLE 4 Species and Strain Distribution of Candida spp. Nosocomial Isolates Number of a

Study Site OR

CA

IA

TX

GA

NY

Species C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C. C.

albicans glabrata tropicalis parapsilosis krusei albicans glabrata tropicalis parapsilosis krusei albicans glabrata tropicalis albicans glabrata parapsilosis albicans glabrata tropicalis parapsilosis lusitaniae guilliermondii albicans glabrata

Isolates

DNA Typesb

72 7 14 64 9 25 3 7 4 2 42 4 3 18 6 2 19 5 4 7 2 2 54 18

30 5 6 36c 5c 24 2 4c 4 1 19c 4 2 10c 5c 2 18c 4c 3c 5c 1 2 18 7

a

Includes only those species in which two or more isolates have been typed. b DNA type determined by a combination of REA and EK analysis. c Two or more isolates from different patients with the same DNA type.

Evidence of possible nosocomial transmission was limited; however, a single DNA type of Candida was shared by $2 patients in 5 of the 6 institutions (Table 5). In most of these instances, there was little or no temporal or spatial relationship between the patients, and thus, it is difficult to interpret the molecular data in the absence of a more detailed epidemiologic investigation. The OR C. parapsilosis isolates were obtained during the investigation of a cluster of four infections occurring in the NICU from November 1994 through January 1995. An additional 6 patients were found to be colonized, and 23 health care workers carried C. parapsilosis on their hands. Molecular typing revealed a large number of different strains (Table 4); however, three small clusters of two patients each, sharing the same strain, were detected (Table 5). Hand cultures of two NICU health care workers were found to be positive with the patient strain of C. parapsilosis coincident with the infection in one of the clusters. In this instance, molecular typing supported the suspicion of nosocomial transmission. A similar situation was noted in the OR SICU, where two colonized patients and three health

M.A. Pfaller et al.

294

FIGURE 1 Representative REA profiles of C. albicans isolates from nine patients hospitalized in the SICU in IA. REA was performed with the restriction enzyme BssHII followed by PFGE. Lanes S contain l phage molecular size standards. Lane

Patient

Date

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

1 1 1 1 1 2 2 2 2 2 3 3 3 4 4 5 6 7 7 7 8 9 9

4/22/94 4/25/94 4/26/94 4/26/94 4/27/94 4/24/94 4/25/94 4/28/94 5/5/94 5/9/94 5/31/94 6/29/94 7/18/94 7/22/94 7/26/94 9/24/94 11/9/94 11/2/94 11/6/94 11/10/94 12/18/94 12/29/94 1/4/95

Source Blood Blood Tissue Tissue Tissue Blood Peritoneal Blood Blood Blood Tissue Catheter Catheter Peritoneal Peritoneal Tissue Peritoneal Blood Blood Blood Blood Peritoneal Peritoneal

fluid

fluid fluid fluid

fluid fluid

DNA type 1 1 1 1 1 2 2 2 2 2 3 3 3 4 4 5 6 7 7 7 7 8 8

care workers were found to share the same strain of C. krusei (Table 5). Two additional patients in this SICU were noted to be fungemic with C. krusei, but these isolates were different from one another and from the other patient and health care worker strains (Berrouane et al. 1996). The only other notable instance of potential nosocomial transmission was the detection of a strain of C. albicans shared by two patients and a health care worker in TX (Table 5). The two patients were hospitalized in the same ICU, and the nurse caring for them was found to be carrying the infecting strain of C. albicans on her hands at the time of infection in one

FIGURE 2 Representative REA profiles of C. albicans isolates from a single patient hospitalized in the NICU in NY. REA was performed with the restriction enzyme SfiI followed by PFGE. Lanes S contain l phage molecular size standards. Lanes 1–20 correspond to the same samples in each gel. All REA and EK profiles are identical. Lane

Date

Source

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

6/2/94 6/2/94 6/3/94 6/3/94 6/3/94 6/3/94 6/5/94 6/7/94 6/9/94 6/9/94 6/10/94 6/17/94 6/16/94 6/21/94 6/22/94 6/23/94 6/29/94 7/13/94 7/21/94 7/26/94

Blood Blood Blood Stool Stool Stool Blood Blood Urine Stool Stool Stool Blood Stool Stool Stool Stool Stool Stool Stool

of the patients. These data document at least three instances of cross-transmission in which patientinfecting strains of Candida spp. were detected on the hands of health care workers and suggest that nosocomial transmission in the ICU setting is not uncommon. The results of this 2-year epidemiological study document considerable variation in the species of Candida causing nosocomial infections in six different medical centers. Infections were caused by several different species of Candida in the SICU setting; however, infections in the NICU were caused almost exclusively by C. albicans and C. parapsilosis. Among the non-albicans species of Candida, C. parapsilosis predominated at two of the study sites. It

National Epidemiology of Mycoses Survey TABLE 5 Study Site

295

NEMIS Study: Analysis of Infected and Colonized Patients with the Same Strain of Candida spp. Species

DNA Type

C. parapsilosis

ORCP-G11

2 patients

ORCP-A11

2 patients 2 HCW

ORCP-M11

2 patients

C. krusei

ORCK-R

2 patients 3 HCW

CA

C. tropicalis

UCCT-21

2 patients

IA

C. albicans

IACA-7G

2 patients

TX

C. albicans

SACA-9I

2 patients

SACA-6F

2 patients 1 HCW

C. glabrata

SACG-3E

2 patients

C. albicans

ECA-3C

2 patients

C. glabrata

ETG-3D

2 patients

C. tropicalis

ECT-1A

2 patients

C. parapsilosis

ECP-3B

2 patients

OR

GA

GA

No. of Patients

was most prominent in the NICU setting and was more common than C. albicans in the OR and TX NICUs. The importance of this organism as an exogenous pathogen was emphasized by the molecular and epidemiologic evidence for nosocomial transmission of a single strain between a health care worker and patients in the NICU of one institution. In contrast to C. parapsilosis, C. glabrata was almost never seen in the NICU, yet was the most common non-albicans species in several of the SICUs in this study. Isolation of this organism has been associated with utilization of azole antifungals (Price et al. 1994; Wingard et al. 1993). Strains of this species requiring elevated azole MICs were observed in the SICU of one institution in this study. The results of this study underscore the diversity in species, strains, and antifungal susceptibility among nosocomial isolates of Candida species. We have confirmed that the vast majority of patients are infected with their own distinct strain of Candida and represent nosocomial infections originating from an

Comment Two colonized (stool) patients in NICU at the same time. One candidemic patient and one colonized patient in NICU at the same time. Hand cultures of two NICU HCW positive coincident with the infection. Two colonized (stool) patients in NICU at the same time. Two colonized (stool) patients in SICU separated by 2 months. Hand cultures of three SICU HCW positive 3–5 months prior to colonized patients. Two infected (wound and pleural fluid) patients in SICU separated by 5 months. Two candidemic patients in SICU separated by 4 weeks (see Figure 1, lanes 18–21). Two candidemic patients in SICU separated by 1 year. One candidemic and one colonized patient in SICU separated by 2 weeks. Hand cultures of one HCW positive coincident with the infection. Two candidemic patients in SICU separated by 4 months. Two candidemic patients in two different units within the same week. Two infected patients in SICU separated by 4 weeks. Two infected patients (peritoneal fluid) in SICU separated by 2 months. Two infected patients (blood and peritoneal fluid) on two different units separated by 2 years.

endogenous source. Although exogenously acquired infection or colonization is much less common, our findings also point to the hands of health care workers as potential reservoirs for nosocomial strains of Candida species, especially C. parapsilosis and C. krusei. The variation in MICs of fluconazole and itraconazole among the different study sites is also noteworthy. Elevated azole MICs attributed to strains of C. glabrata and C. albicans from patients on azole therapy bears careful watching, and should be compared to the institution’s usage of specific antifungal agents. Continued surveillance will be necessary to clarify further the true incidence and clinical importance of the emergence of new strains and species of Candida in the ICU setting.

This study was supported by a grant from Roerig-Pfizer. The authors thank Kay Meyer for excellent secretarial support.

296

M.A. Pfaller et al.

REFERENCES Banerjee SN, Emori TG, Culver DH, Gaynes RP, Jarvis WR, Horan T, Edwards JR, Tolson J, Henderson T, Martone WJ (1991) Secular trends in nosocomial primary blood stream infections in the United States, 1980–1989. Am J Med 91(suppl 3B):86S–89S. Beck-Sague CM, Jarvis WR (1993) National Nosocomial Infections Surveillance System. Secular trends in the epidemiology of nosocomial fungal infections in the United States, 1980–1990. J Infect Dis 167:1247–1251. Berrouane YF, Hollis RJ, Pfaller MA (1996) Strain variation and antifungal susceptibilities among isolates of Candida krusei. J Clin Microbiol 34:1856–1858. Emori TG, Gaynes RP (1993) An overview of nosocomial infections, including the role of the microbiology laboratory. Clin Microbiol Rev 6:428–442. Fisher-Hoch SP, Hutwagner L (1995) Opportunistic candidiasis: an epidemic of the 1980s. Clin Infect Dis 21:897– 904. Fridkin SK, Jarvis WR (1996) Epidemiology of nosocomial fungal infections. Clin Microbiol Rev 9:499–511. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM (1988). CDC definitions for nosocomial infections, 1988. Am J Infect Cont 16:128–140. Hollis RJ, Doebbeling BN, Pfaller MA (1995) Restriction endonuclease analysis of genomic DNA from Candida species. In: Molecular Biology of Pathogenic Fungi: A Laboratory Manual. Eds, Maresca B, Kobayashi GS. New York: Springer-Verlag. Jarvis WR (1995) Epidemiology of nosocomial fungal infections, with emphasis on Candida species. Clin Infect Dis 20:1526–1530. National Committee for Clinical Laboratory Standards (1997) Reference Method for Broth Dilution Antifungal Susceptibility Testing for Yeasts: Approved Standard. NCCLS Document 27-A. Wayne, PA: NCCLS. Nguyen MH, Peacock JE, Morris AJ, Tanner DC, Nguyen ML, Snydman DR, Wagener MM, Rinaldi MG, Yu VL (1996) The changing face of candidemia: emergence of non-Candida albicans species and antifungal resistance. Am J Med 100:617–623. Pfaller MA (1995a) Epidemiology of candidiasis. J Hosp Infect 30(suppl):329–338. Pfaller MA (1995b) Epidemiology of fungal infections: cur-

rent perspectives and future directions. Clin Infect Dis 20:1525. Pfaller MA (1996) Nosocomial candidiasis: emerging species, reservoirs, and modes of transmission. Clin Infect Dis 22(suppl 2):S89–S94. Pfaller MA, Bale M, Buschelman B, Lancaster M, EspinelIngroff A, Rex JH, Rinaldi MG, Cooper CR, McGinnis MR (1995) Quality control guidelines for National Committee for Clinical Laboratory Standards recommended broth macrodilution testing of amphotericin B, fluconazole, and flucytosine. J Clin Microbiol 33:1104–1107. Pfaller MA, Cabezudo I, Koontz F, Bale M, and Gingrich R (1987) Predictive value of surveillance cultures for systemic infection due to Candida species. Eur J Clin Microbiol 6:628–633. Pfaller MA, Hollis RJ, Sader HS (1994) Chromosomal restriction fragment analysis by pulsed-field gel electrophoresis. In: Clinical Microbiology Procedures Handbook, Suppl 1. Ed, Isenberg HD. Washington, DC: American Society for Microbiology. Pinner RW, Teutsch SM, Simonsen L, Klug LA, Graber JM, Clarke MJ, Berkelman RL (1996) Trends in infectious diseases mortality in the United States. J Am Med Assoc 275:189–193. Pittet D, Wenzel RP (1995) Nosocomial blood stream infections. Arch Intern Med 155:1177–1184. Strausbaugh LJ, Sewell DL, Tjoelker RC, Heitzman T, Webster T, Ward TT, Pfaller MA (1996) Comparison of three methods for recovery of yeasts from hands of healthcare workers. J Clin Microbiol 34:471–473. Warren NG, Hazen KC (1995) Candida, Cryptococcus, and other yeasts of medical importance. In: Manual of Clinical Microbiology. Eds, Murray PR, Baron EJ, Pfaller MA, Tenover FC, Yolken RH. Washington, DC: American Society for Microbiology, pp. 723–737. Wenzel RP (1995) Nosocomial candidemia: Risk factors and attributable mortality. Clin Infect Dis 20:1531–1534. Wey SB, Mori M, Pfaller MA, Woolson RF, Wenzel RP (1988) Hospital-acquired candidemia: The attributable mortality and excess length of stay. Arch Intern Med 148:2642–2645. Wey SB, Mori M, Pfaller MA, Woolson RF, Wenzel RP (1989) Risk factors for hospital-acquired candidemia. Arch Intern Med 149:2349–2353.