Supported by an unrestricted educational grant from Pfizer, Inc.
The most common fungal pathogens associated with invasive disease in humans are opportunistic yeasts (e.g., Candida albicans) or filamentous fungi (e.g., Aspergillus spp.). Fungi previously thought to be nonpathogenic for humans or only sporadically associated with human disease, such as Candida (except albicans), Fusarium, Trichosporon, and Malassezia spp., are emerging as important nosocomial fungal pathogens (Table 1). These pathogens are associated with increasing morbidity and mortality. The emergence of these organisms and antifungal-resistant fungi, especially those that are azole-resistant, poses an important challenge to the clinician.
National Nosocomial Infections Surveillance (NNIS) System data show that fungi account for 9% of all nosocomial infections (Table 2). From 1980 through 1990, the nosocomial fungal infection rate increased from 2.0 to 3.8 infections per 1,000 patient discharges. The most common fungi reported were Candida spp. (85.6%), followed by Aspergillus spp. (1.3%). C. albicans accounted for 76% of all Candida spp. infections. Other fungal pathogens (e.g., Malassezia, Trichosporon, Fusarium, and Acremonium) represented 11% of the nosocomial fungal pathogens. Increasing secular trends in nosocomial fungal infection rates were seen for urinary tract, and bloodstream infections and pneumonia in the intensive care units (ICUs) reporting to the NNIS System from 1986 through 1996 (Fig. 1). The crude mortality from opportunistic fungal infections exceeds 50% in most studies and has been reported as high as 95% in bone marrow transplant recipients with Aspergillus spp. infection. The attributable mortality for patients with candidemia has been estimated at 38%.
Risk factors for opportunistic fungal infections include receipt of antimicrobial agents, chemotherapy, presence of indwelling intravascular catheters, neutropenia, prior hemodialysis, or previous fungal colonization. Most of these factors are very common in hospitalized patients, especially in high-risk areas such as ICUs or oncology wards.
Species identification and molecular typing have become critical elements of nosocomial fungal outbreak investigations.
Although approximately 200 Candida spp. have been identified, only a few have been associated with human disease. Candida spp. are ubiquitous and found on many plants and as normal endogenous flora in the gastrointestinal (GI) tract of mammals and mucocutaneous membranes of humans. The species most commonly isolated from the human GI tract is C. albicans but other non-albicans Candida spp.--most commonly C. tropicalis, C. parapsilosis, and Torulopsis glabrata--have been isolated.
Candida spp. may cause a wide variety of diseases in humans, ranging from oral thrush to disseminated invasive disease. Bloodstream infection is the most frequent nosocomial infection caused by Candida spp.
The diagnosis of invasive candidiasis remains a clinical challenge requiring a high index of clinical suspicion and use of cultures, diagnostic imaging techniques, and histopathologic studies. Since Candida spp. may take several weeks to grow in culture, identification of the species is often delayed. Also, sensitivity of routine blood cultures is low, although the use of lysis centrifugation or biphasic blood culture systems appears to enhance recovery. New molecular diagnostic techniques (e.g., polymerase chain reaction [PCR]) are being developed and evaluated and may improve diagnosis and facilitate early treatment of infected patients.
Nosocomial C. albicans outbreaks have been associated with extrinsically contaminated products and with transmission from patient to patient and healthcare worker (HCW) to patient. The hands of HCWs can serve as an important source of infection. The first choice therapeutic agent for treating uncomplicated candidemia due to C. albicans is amphotericin B (AmB) or fluconazole.
Over the past 2 decades, there has been a change in the distribution of Candida spp. causing nosocomial infection. Numerous reports have documented the emergence of C. tropicalis, C. glabrata, C. parapsilosis, C. krusei, and C. lusitaniae. This may be related to the increased use of antifungal agents, particularly as prophylactic or empiric agents. Further studies are needed to clarify the relationship between the use of antifungal agents and the emergence of resistant or unusual strains.
C. parapsilosis has been associated with endemic and epidemic nosocomial infections traced to total parenteral nutrition or intravascular devices. C. parapsilosis can be isolated from HCWs' hands, particularly of those working in neonatal ICUs. This species also colonizes the GI tract of humans and there is evidence that GI colonization in neonates may precede infection. The crude mortality rate associated with C. parapsilosis bloodstream infection has been estimated to be 30%. C. parapsilosis is usually susceptible to AmB and fluconazole.
C. tropicalis appears to be more virulent than C. albicans in patients with hematologic malignancy and disseminated infection is associated with high mortality rates. Most infections appear to originate from the patients' endogenous microflora, although 1 outbreak of fungemia in a neonatal ICU was associated with receipt of parenteral nutrition and antimicrobial agents. An outbreak of C. tropicalis sternal wound infections following cardiac surgery was traced to a colonized scrub nurse. Most C. tropicalis are susceptible to AmB, flucytosine, and triazoles. As with other fungi, flucytosine should not be used as monotherapy because of development of resistance.
C. krusei causes fungemia and endophthalmitis. Endogenous spread from the GI tract is the main mechanism of infection, particularly in patients with hematologic malignancy. Several reports have described an increase in C. krusei colonization and infection in granulocytopenic patients associated with use of fluconazole as prophylaxis. Molecular studies suggest exogenous acquisition by cross-infection. Since C. krusei is resistant to fluconazole, therapeutic options include AmB and flucytosine.
Reports of C. glabrata infections have increased over the past decade. C. glabrata infections appear to be more common in patients with solid tumors and nononcologic disorders. Colonization at multiple sites usually precedes infection and portals of entry include the respiratory tract, surgical wounds, or the genitourinary tract. This species may be emerging secondary to increasing device use and treatment with prophylactic azole agents to which this organism is intermediately resistant. Strains may be susceptible to AmB and flucytosine.
C. lusitaniae is an important cause of nosocomial infection among immunocompromised hosts. It is an endogenous pathogen and nosocomial transmission between patients can occur. Initial resistance or rapid development of resistance to AmB is characteristic of this species. Treatment of fungemia requires higher doses of AmB and/or AmB in combination with flucytosine or fluconazole. Most strains are susceptible to azoles.
Until the epidemiology and pathogenesis of Candida spp. is better understood, prevention efforts should be targeted to patients at high risk. Infection control guidelines to interrupt or prevent the transmission of infection are crucial and involve judicious use of antimicrobials, HCW handwashing, and adherence to aseptic technique and isolation precautions.
Two Malassezia spp. (M. furfur and M. pachydermatis) have been associated with most human disease. These lipophilic and saprophytic organisms are human skin commensals and colonize oily areas of the skin, especially the scalp, back, and chest.
M. furfur is the etiologic agent of pityriasis versicolor. Reports of invasive infection, including catheter-associated sepsis and peritonitis associated with peritoneal dialysis, are increasingly common. M. furfur commonly colonizes neonates in neonatal ICUs. Such colonization is associated with prematurity, duration of neonatal ICU stay, use of occlusive dressings, and prolonged antibiotic therapy. Invasive disease usually involves low birthweight neonates and sporadically other highly immunocompromised populations. M. furfur is dependent on exogenous lipid. A common factor associated with fungemia in susceptible patients is exposure to intravenous lipids and parenteral nutrition. M. furfur can also be transmitted from patient to patient on the hands of HCWs.
Clinical features of disseminated M. furfur infection in neonates include fever, bradycardia, respiratory distress, hepatosplenomegaly, lethargy, or seizures. Approximately 50% of neonates and 30% of adults with disseminated disease have had interstitial pneumonitis. Treatment includes cessation of lipid infusion, catheter removal, and selected use of systemic antifungal agents (AmB or imidazoles).
Recently there have been several reports of M. pachydermatis outbreaks in infants in neonatal ICUs. In 1 outbreak, the pathogen was probably introduced into the neonatal ICU on HCWs' hands after the HCWs were colonized from their pet dogs. Patients presented with bloodstream infection, urinary tract infection, or meningitis. Diagnosis of Malassezia spp. requires culture of the body sites potentially involved. Laboratory personnel should be alerted because special media or use of oil overlay may be necessary. Malassezia spp. usually are susceptible to AmB.
Trichosporon beigelii (previously T. cutaneum) is the cause of superficial scalp, beard, and axillary and pubic hair infection, called white piedra. This fungus can be part of the normal skin flora, stool, or urine. Invasive diseases, including bloodstream infections, endocarditis, and peritonitis, have been reported--particularly in severely immunocompromised patients. The organism is rarely recovered from the hospital environment. Patients with disseminated infection are acutely ill with fever and symptoms of organic dysfunction such as renal, cardiac, and pulmonary dysfunction. Diagnosis of disseminated disease is dependent on blood culture or histopathologic examination of tissue.
T. beigelii may cause false-positive Cryptococcus neoformans serum latex agglutination tests. Infection is associated with high morbidity and mortality and is difficult to treat, with frequent resistance to azoles and AmB. The optimal antifungal therapy for these infections is unclear. Some suggest empiric use of AmB and flucytosine or rifampin before susceptibility testing results are available.
In the United States, Aspergillus spp. are the second most common cause of hospital-acquired fungal infection. Several Aspergillus spp. have been associated with human infections. A. fumigatus, A. flavus, and A. terreus are the most common.
Risk factors for invasive aspergillosis include granulocytopenia, neutropenia, and use of corticosteroids or cytotoxic drugs. Immunocompromised patients exposed to sources contaminated with Aspergillus spp. (e.g., air on construction sites) may become colonized and subsequently infected. Nosocomial Aspergillus spp. outbreaks have been related to contaminated air or ventilation systems.
Clinical disease is predominantly respiratory followed, in some patients, by dissemination. It may present as allergic bronchopulmonary aspergillosis, aspergilloma, Aspergillus pneumonia, or invasive pulmonary aspergillosis. Endocarditis, postoperative infections, and infections associated with peritoneal catheters have been reported. Otomycosis, sinus infection, and eye infection with Aspergillus spp. can occur. In immunosuppressed patients, there can be esophageal or GI ulceration, necrotizing skin ulcers, renal involvement, bone lesions, and central nervous system infection.
Diagnosis of invasive aspergillosis remains a clinical challenge. Premortem blood culture sensitivity is low so definite diagnosis requires histopathologic evidence of Aspergillus spp. invasion. Cultures obtained from the respiratory tract may be used in conjunction with clinical diagnosis, diagnostic imaging techniques, and serology but colonization may exist without disease. Furthermore, nasal cultures of immunocompromised patients are not predictive of invasive disease.
Treatment includes antifungal therapy with AmB in conjunction with surgical excision in some patients.
The most common pathogenic Paecilomyces spp. are P. variotii and P. lilacinus. Paecilomyces spp. are found in the soil. They have contaminated sterile solutions or clinical specimens. P. lilacinus has been associated with eye infections, including keratitis, endophthalmitis, corneal ulcer, and orbital granulomas; soft tissue infection; noninvasive sinusitis; and skin infection. Nosocomial transmission has been associated with intrinsic contamination of a neutralizing solution for synthetic intraocular lenses and with extrinsic contamination of a skin lotion.
P. variotii has been associated with endocarditis following valve replacement, peritonitis in patients undergoing peritoneal dialysis, pyelonephritis, sinusitis, pulmonary infection, cellulitis, and fungemia in both immunocompetent and immunosuppressed patients.
Identification of the species is important because P. lilacinus is resistant to AmB and, by in-vitro testing, susceptible to azoles. P. variotii is susceptible to AmB and flucytosine. However, regardless of infecting species, clinical outcome has been poor.
Acremonium spp. are common in the environment and include approximately 100 species. Most infections caused by Acremonium spp. are mycetomas. Reports of disseminated and invasive infections are rare and the mechanism of infection is thought to be secondary to increased host susceptibility and prior colonization.
Nosocomial transmission is unusual. In 1 outbreak of endophthalmitis in patients following cataract surgery, the reservoir was a contaminated humidifier, which was distal to the high efficiency particulate air filter in the surgical center ventilation system.
Treatment regimens are not well defined, although it has been suggested that AmB is most active for treating disease. Several patients with endophthalmitis have been treated with vitrectomy and systemic antifungals.
Infection with Zygomycetes is associated with high mortality, particularly in severely ill patients. The fungus usually enters the body through the respiratory tract or is introduced directly onto abraded skin. Primary infection sites are the sinus cavities, lungs, skin, GI tract, and central nervous system.
Risk factors include hematologic malignancies, myelosuppression, renal failure, diabetes mellitus, receipt of antimicrobial agents, severe underlying disease, or exposure to hospital construction activity. Reservoirs and mechanisms of transmission are similar to those of Aspergillus spp.
Outbreaks of primary cutaneous mucormycosis have been associated with use of nonsterile or contaminated adhesive tape or elasticized surgical bandages. Diagnosis is dependent on biopsy results. Therapy requires use of AmB and surgical debridement.
The lack of specific and rapid diagnostic tests for fungal infections is one of the major impediments to successful management of infected patients. Numerous studies have shown that delay in appropriate therapy is associated with increased morbidity and mortality. Early definitive diagnosis is imperative. Therefore, nosocomial fungal infections should be considered in the differential diagnosis of high-risk patients and appropriate clinical specimens for testing should be promptly obtained.
Molecular methods for the rapid diagnosis of fungal infection provide hope for the future. Antigen detection systems, if specific and sensitive, should also facilitate more rapid diagnosis of invasive disease and earlier institution of directed therapy. Currently being evaluated are several such tests, including the galactomannan enzyme immunoelectrophoresis assay (EIA) for rapid diagnosis of invasive aspergillosis and PCR-EIA to detect Candida spp. in blood.
Being aware of the epidemiology and pathogenesis of fungal infections is crucial to minimize the risk of acquiring a nosocomial fungal infection. Polyenes (primarily AmB), flucytosine, and the azoles are the main antifungals available for prophylaxis and treatment of fungal infection.
New lipid formulations of AmB, now available for clinical use, employ liposomes as a target drug delivery system for AmB in an attempt to attenuate possible AmB adverse effects and increase its therapeutic potential. Three liposomal formulations of AmB are commercially available. Published studies evaluating these formulations show decreased nephrotoxicity but no improvement in outcome compared to conventional AmB in the treatment of documented infections. AmBisome (L-AmB) is an available lipid formulation of AmB. When used for empirical treatment of febrile neutropenia, it reduced the incidence of proven emergent fungal infections but did not improve short-term survival rates when compared to conventional AmB. High cost is the main impediment to the widespread use of these lipid formulations. Daily treatment cost for these agents ranges from 14 to 54 times the cost of conventional AmB therapy, depending on the agent used.
New azole formulations are being evaluated for prophylaxis and therapy of fungal infections. At present, there are insufficient data to support the use of 1 agent over another.
Although there have been reports of emerging resistance to azoles and AmB, the extent is unknown. Recently, susceptibility testing standards for yeasts have been developed by the National Committee for Clinical Laboratory Standards. Such standards are being developed for filamentous fungi. Few data exist to correlate these standards and clinical outcomes. As antifungal susceptibility testing becomes more widespread, the correlation between in vitro resistance and in vivo clinical failure can be determined.
With concerns about the emergence of antifungal resistance, it is essential that prophylactic and empiric antifungal therapy be based on appropriately designed clinical studies, particularly in high-risk patients. Further studies are needed to define optimal approaches.
Prevention and control of nosocomial fungal infections require application of current prevention recommendations, a high index of suspicion, application of the most sensitive and specific appropriate diagnostics, and specifically directed use of antifungal agents. Current research in each of these areas should enhance prevention measures and diagnostics and facilitate earlier treatment, which will improve patient outcomes.
Cristiana M. Toscano, MD
and
William R. Jarvis, MD
Investigation and Prevention Branch
Hospital Infections Program
Centers for Disease Control and Prevention
Atlanta, Georgia