Fungal infections

Epidemiology

Candida spp. remain among the 10 most common pathogens of healthcare–associated infections, but are the most common cause of nosocomial bloodstream infections (BSIs). Candida spp. have consistently caused a substantial disease burden for more than a decade. However, over the past decade, population-based studies have noted a decline in overall incidence, which likely reflects improvements in healthcare delivery. Worldwide, Candida spp. are the third most frequent cause of infection in ICUs. Traditionally, ICUs have a higher incidence of Candida BSIs than medical and surgical wards, although recent data suggest an increasing prevalence of such infections in non-ICU patients. ^, Although prior data had suggested the frequency of Candida BSIs among ICU patients had declined, estimates from national secondary databases and population-based studies suggest the disease burden may be shifting from the ICU to the general hospital population. ^,

Globally, C. albicans remains the most common invasive Candida spp. However, decreasing trends in the isolation of this species over time have been observed in the ICU and non-ICU setting. ^{, ,} The isolation of C. albicans and C. parapsilosis among neonatal ICU patients and the prevalence of C. glabrata infections among adults have been widely appreciated. ^, C. albicans is responsible for approximately 45% of episodes of candidemia. ^, The incidence of infection caused by a particular Candida spp. varies considerably by the clinical service on which the patient is hospitalized. However, in general, C. albicans is the primary fungal pathogen in the ICU setting and is followed by C. glabrata, C. parapsilosis, C. tropicalis, C. krusei, and other Candida spp. (i.e., C. guilliermondii, C. lusitaniae ). ^, This rank order varies little across infection site, but it may vary with age, underlying disease, or local epidemiology. ^{, , ,} Surveillance data have noted that candidemia in neonatal ICUs is predominantly the result of C. albicans and C. parapsilosis and rarely the result of C. glabrata or other Candida spp. ^{, , ,} Surveillance studies have demonstrated that BSI caused by C. albicans occurs less frequently with increasing age. ^{, , ,} In contrast, C. glabrata is rarely isolated among infants and children but is more frequently found with increasing patient age. ^{, , ,}

In 2009 a new multidrug-resistant species, Candida auris, first emerged and subsequently, infections caused by this species have spread across the globe. Compared with other known non- albicans species, the emergence of C. auris has fostered significant concern because of its rapid global spread and its multidrug-resistance profile. Genomic sequencing has identified at least four major phylogenetically distinct populations (clades), each of which geographically cluster. Widespread genetic variation exists between each clade, but within each population, there is minimal genetic diversity. Interclade genetic variation leads to differences in pathogenicity, biochemical characteristics, and antifungal susceptibility. For example, strains comprising the East Asian clade are associated with noninvasive infections and appear to be more susceptible to antifungal therapy than isolates of the other three clades. C. auris is phenotypically similar to other Candida spp., but phylogenetically it is most closely related to Candida haemulonii . Historically, C. auris has often been misidentified by commercial assays, so prevalence estimates for this pathogen and the frequency with which it causes infections in the ICU are difficult to ascertain. Nonetheless, before its emergence in 2009, it appears the prevalence of C. auris was exceedingly rare. ^, C. auris infections occur in ICUs among critically ill patients of all ages, those undergoing invasive procedures, or those with serious underlying conditions that affect their host defenses. ^,

C. albicans is part of one’s microbiota. Infections, including BSIs caused by most Candida spp., particularly C. albicans , arise endogenously from the gastrointestinal mucosa, skin, and urinary tract. Invasive Candida infections occur when alteration of the patient’s microbiota leads to overgrowth of yeast which, in the presence compromised skin or gastrointestinal mucosa integrity, translocates from its commensal environment to the bloodstream. Candida spp., including C. albicans, may also be transmitted exogenously in ICU settings. ^, Exogenous transmission of non- albicans Candida spp. through indirect contact with the ICU environment occurs commonly. For example, C. parapsilosis is known for its ability to form biofilms on catheters and inert devices, which enables it to persist in the nosocomial environment. Moreover, it is spread throughout the hospital through hand carriage by healthcare workers. Infections caused by C. auris can arise from colonization, travel to endemic regions, and making contact with the hospital environment. C. auris frequently colonizes the axilla and groin, but it can also be isolated from a variety of body sites. Some C. auris isolates demonstrate the capacity to form biofilms and can survive on inert surfaces. Estimates suggest that acquisition from a contaminated environment or colonized patient can occur in as little as 4 hours, and invasive infections can reportedly develop within a couple days to weeks of admission to an ICU. ^{, ,}

Detection and isolation

Candida BSIs are often difficult to detect. Symptomatically, BSIs caused by Candida spp. are indistinguishable from BSIs of bacterial etiology. Isolating Candida spp. from blood is challenging because they are cleared rapidly from the circulation by several organs, particularly the liver. Blood and deep-seated tissue cultures yield positive results approximately 50% of the time, have slow turnaround, and may not manifest until late in the disease course. Deep-seated tissue cultures also must often be collected by invasive techniques, which depending on comorbidities, may not be possible. Despite these limitations, these culture-based methods remain the “gold standard” in the diagnosis of candidemia and other forms of invasive candidiasis. Moreover, the ability of automated blood culture systems to recover Candida spp. has continued to improve, and nonculture methods are available that, when used in with culture-based methods, can identify more infected patients earlier.

The biochemical characteristics associated with C. auris are difficult to distinguish from other species and can vary across clades. Consequently, routinely used, commercially available identification methods based on phenotypic assimilation/fermentation tests often misidentified C. auris as a variety of other Candida spp. Much of the problem has been related to the lack of representative isolates, or clade diversity, in the databases of most commercially available biochemical identification systems. This gap should close over time, with system database updates as information and experience with this emerging pathogen continue to grow. Molecular methods have some advantages over biochemical methods, but such systems are still reliant upon accurate and up-to-date databases for comparison to the analyzed sample. Still, molecular diagnostic methods developed to enable accurate and timely diagnosis should be employed as needed to complement biochemical methods.

Mortality

Estimating the mortality caused by Candida infections is challenging, particularly in critically ill patients, and is dependent on the underlying conditions and specific patient population. Nonetheless, earlier studies continually demonstrated Candida BSIs carry a relatively poor prognosis. Candida spp. isolated from the blood have been identified as an independent predictor of mortality. Whether expressed as part of all-cause mortality or as crude mortality estimates, Candida BSIs are associated with high rates of death. Estimates for mortality rates associated with Candida BSIs hospital-wide and in the ICU range from 30% to 40%, with rates as high as 72% in certain countries. ^, Reported mortality rates associated with the emerging pathogen C. auris have varied substantially; whether species-related differences in mortality rates exist is difficult to quantify. With more reporting, it is likely that mortality rates associated with invasive infections caused by C. auris will be similar to that reported with other Candida spp.

Mortality rates associated with candidemia and other forms of invasive candidiasis hospital-wide and in the ICU, though variable, remain quite high despite the advent of potent and safer anti- Candida antifungal therapy. ^{, , ,} Inadequate treatment may be a reason why mortality has not improved despite the availability of potent and safe antifungal therapy. Inadequate therapy resulting from delays in administration, treatment with an agent to which the organism is resistant, inadequate dosing or treatment duration, or failure to recognize and treat candidemia all contribute to the mortality associated with Candida BSI. Delaying initiation of adequate antifungal therapy even 12–48 hours is independently associated with mortality in candidemia patients. ^{, , ,}

Candidemia produces significant morbidity and increases the length of hospital stay. ^, Given the severity of illness associated with this infection, the added length of stay uses significant healthcare resources.

Risk factors

Among critically ill patients, risk factors for Candida infections are well known. ^, Broad-spectrum antimicrobial use, colonization, indwelling vascular catheters, extremes of age, and hemodialysis have been consistently identified as independent risk factors for Candida BSIs. ^, Many of these diverse risk factors are commonly present and unavoidable in ICUs. The ICU itself provides an ideal environment for transmission of Candida spp., including C. auris, among patients; thus it is not surprising that prolonged ICU stay has been identified as an independent risk factor. ^, For these reasons, combined with the high mortality rates associated with candidemia and invasive candidiasis, a variety of predictive models to identify patients most at risk for infection have been developed using clinical and/or microbiologic variables. Early efforts using only microbiologic data gained through surveillance cultures focused on the relationship between colonization and infection and sought to quantify it through a so-called “colonization index.” The use of resource-intensive surveillance cultures and a lack robust validation studies limit the usefulness of such approaches. Predictive models based solely on clinical variables generally demonstrate poor positive predictive values, and consequently have high negative predictive values (i.e., they are better at identifying those who are unlikely to benefit from antifungal therapy). A model designed using microbiologic data and clinical variables in non-neutropenic patients colonized with Candida spp. with a minimum length of ICU stay of 7 days. Performed slightly better at identifying at-risk patients and may be useful for recognizing the need for prompt initiation of antifungal therapy. Although these models may have some utility in curbing inappropriate antifungal therapy, they are somewhat complicated to apply, and certain components of individual prediction rules may be impractical to collect or determine. ^, Given the diversity of risks associated with developing infection, caution should be used when employing predictive models for candidemia and invasive candidiasis in critically ill patients. Clinicians should also be aware of the patient population and the clinical conditions used to derive a given predictive model and understand it may not be generalizable to their patient.

Invasive aspergillosis in critically ill patients with hematologic malignancies

In contrast to Candida spp., the burden of infection caused by Aspergillus spp. is small. ^, Aspergillus spp. cause infection in critically ill populations immunocompromised by burns, cytotoxic chemotherapy, prolonged corticosteroid therapy, malignancy, leukemia, SOT or HSCT, and other congenital or acquired immunodeficiencies. Aspergillus spp. are ubiquitous environmental molds. Although several hundred species of Aspergillus have been described, relatively few are known to cause disease in humans. Most Aspergillus infections are acquired exogenously via inhalation. In the absence of an effective immune response, airborne conidia invade sinus or lung vasculature. Although the lung is the most common site of invasive aspergillosis, Aspergillus spp. also demonstrate tropism for cutaneous, central nervous system (CNS), bone, and cardiac vasculature.

The incidence of invasive aspergillosis in immunocompromised patients varies among specific populations. Among patients with hematologic malignancies, those with acute myelogenous leukemia have the highest incidence of invasive aspergillosis. Patients in ICUs are at increased risk, and susceptibility in general depends on the use of immunosuppressants, structural lung damage, and genetic predisposition. Like patients with leukemia, patients undergoing HSCT are at high risk for invasive aspergillosis. The incidence of invasive aspergillosis varies depending on transplant type but not the type of conditioning regimen (myeloablative vs. nonmyeloablative). The incidence is higher among allogeneic HSCT recipients than among autologous HSCT recipients. In the HSCT population, whether the incidence of invasive aspergillosis is truly increasing or decreasing is difficult to ascertain, because the rate of autopsy continues to decline. The incidence of invasive aspergillosis among SOT is highest among lung transplant recipients and lowest among renal transplant recipients. Patients receiving HSCT or SOT can develop invasive aspergillosis shortly (within 40 days) after transplantation, but typically it occurs late post-HSCT (>40–100 days) or SOT (>90 days).

In patients with acute leukemia or in HSCT recipients, prolonged neutropenia after cytotoxic chemotherapy or HSCT is the primary risk for early invasive aspergillosis. Risk factors associated with invasive aspergillosis in HSCT and SOT recipients vary with time after the transplant. However, in general, risks early in the transplant process are associated with transplant-related factors (underlying disease, neutropenia, type of transplant), biologic factors (hyperglycemia, iron overload), and extrinsic factors (excluding spores from the environment, air filtration). In contrast, risks for invasive aspergillosis occurring later in the transplant process include transplant complications (acute graft-versus-host disease [GVHD] [grade ≥3] and high-dose corticosteroid therapy).

Lesions associated with invasive pulmonary aspergillosis evolve over a period of weeks. Computed tomography (CT) findings, especially nodular infiltrates with “halo sign,” are strongly suggestive of invasive aspergillosis and infection from other angioinvasive fungi in immunocompromised patients. Moreover, this finding is associated with significantly improved response and survival if antifungal therapy is initiated shortly upon detection of this sign of infection.

Recent diagnostic efforts have focused on detecting non–culture-based serum markers (e.g., galactomannan test, 1,3-β- d -glucan, polymerase chain reaction [PCR]). Galactomannan is a cell wall constituent of Aspergillus spp. that can be detected in the serum during invasive infection. The test is specific for invasive aspergillosis and is commercially available as a sandwich enzyme immunoassay (enzyme-linked immunosorbent assay [ELISA]) that detects circulating galactomannan. The values from this test have been shown to strongly correlate with the clinical outcome of patients with invasive aspergillosis. Because 1,3-β- d -glucan is a cell wall component of many fungal pathogens, it can be detected by colorimetric detection assays. Although the test is highly sensitive, the presence of 1,3-β- d -glucan in the serum is not specific for any fungi. Using both of these non–culture-based serum markers may improve the ability to diagnose invasive aspergillosis in high-risk populations and could lead to earlier diagnosis or improved monitoring of the success of antifungal therapy. ^, CT-guided biopsy has a high diagnostic yield, and samples should undergo both histopathologic and cultural evaluation. The combination of radiologic, serologic, cultural, histopathologic, and clinical data may ultimately improve the diagnosis of invasive aspergillosis and speed up initiation of appropriate antifungal therapy.

Miscellaneous pathogens in critically ill patients with hematologic malignancies

Candida and Aspergillus spp. are the primary fungal pathogens in critically ill patients with hematologic malignancies. However, other pathogens such as Mucorales, Fusarium, Lomentospora, Scedosporium, and other orphan but emerging mold diseases are increasing in frequency. Each of these less common organisms has clinical characteristics or tissue tropism. In addition, they are often less susceptible than Aspergillus spp. to systemic antifungal agents. Consequently, infections caused by these pathogens are associated with high mortality. Of these, the Mucorales are the most common among critically ill patients. These angioinvasive pathogens are acquired through inhalation and produce a necrotic infection with the highest morbidity and mortality. Rhinocerebral, paranasal, pulmonary with or without transdiaphragmatic extension, cutaneous, and gastrointestinal infections are common manifestations of mucormycosis. Common risks are diabetic ketoacidosis, immunosuppression, organ transplantation, traumatic skin damage, and a prolonged ICU stay. ^, Pulmonary and disseminated infection mostly affects patients with hematologic malignancy; rhinocerebral and paranasal mucormycosis is predominant in patients with uncontrolled diabetes. ^, The “reversed halo” sign in CT is indicative of pulmonary mucormycosis; however, diagnosis should be enforced by cultural and histopathologic work-up of biopsies. Management of mucormycosis, including the important role of surgical debridement, is detailed in the global guideline for the diagnosis and management of mucormycosis.

Amphotericin B deoxycholate

Amphotericin B deoxycholate (AmB-d), a polyene antifungal agent, disrupts biologic membranes, thereby increasing their permeability. AmB-d also stimulates the release of cytokines, which causes arteriolar vasoconstriction in the renal vasculature. ^,