Invasive Fungal Infections in Acute Leukemia

Abstract

Invasive fungal infection (IFI) is among the leading causes for morbidity, mortality, and economic burden for patients with acute leukemia. In the past few decades, the incidence of IFI has increased dramatically. The certainty of diagnosis of IFI is based on host factors, clinical evidence, and microbiological examination. Advancement in molecular diagnostic modalities (e.g. non-culture-based serum biomarkers such as β-glucan or galactomannan assays) and high-resolution radiological imaging has improved our diagnostic approach. The early use of these diagnostic tests assists in the early initiation of preemptive therapy. Nonetheless, the complexity of IFI in patients with leukemia and the limitations of these diagnostic tools still mandate astute clinical acumen. Its management has been further complicated by the increasing frequency of infection by non-Aspergillus molds (e.g. zygomycosis) and the emergence of drug-resistant fungal pathogens. In addition, even though the antifungal armamentarium has expanded rapidly in the past few decades, the associated mortality remains high. The decision to initiate antifungal treatment and the choice of anti-fungal therapy requires careful consideration of several factors (e.g. risk stratification, local fungal epidemiologic patterns, concomitant comorbidities, drug-drug interactions, prior history of antifungal use, overall cost, and the pharmacologic profile of the antifungal agents). In order to optimize our diagnostic and therapeutic management of IFI in patients with acute leukemia, further basic research and clinical trials are desperately needed.

Epidemiology

Patients with acute leukemia (AL) are considered a population at high risk for developing an invasive fungal infection (IFI) [Bow, 2008; Chamilos et al. 2006]. In the United States, the incidence of fungal sepsis increased by approximately 200% between 1979 and 2000 [Martin et al. 2003]. This increase is believed to be secondary to the extended survival of AL patients as well as to advances in supportive care, improved control of bacterial infections [Leventakos et al. 2010], and the use of intensive chemotherapy [Sung et al. 2009], immunotherapy regimens, and hematopoietic stem cell transplantation (HSCT) [Bow, 2008].

IFI is a major cause of morbidity and mortality in patients with AL. In a recent study, the cumulative probability of developing IFI after a diagnosis of AL was 11.1% at 100 days [Hammond et al. 2010]. Patients undergoing treatment for hematologic malignancies have an estimated causespecific mortality due to IFI of 35% [Auberger et al. 2008]. Furthermore, IFI appears to be underdiagnosed ante mortem: one recent report identified IFI in 314 of 1017 (31%) autopsies of patients with hematologic malignancies, of which only 25% had been identified ante mortem [Chamilos et al. 2006].

The increase in the incidence of IFI has resulted in a substantial increase in the length of hospital stays and healthcare expenditures [Morgan et al. 2005; Dasbach et al. 2000]. As an example, compared with cancer patients without aspergillosis, cancer patients with this disease stayed in the hospital for an average of 26 more days (33 vs. 7 days), incurred US $115,262 more in total costs on average, and had four times the mortality rate during hospitalization (31% vs. 7%) [Dasbach et al. 2000].

Despite these dire statistics, the death rate due to IFI has dropped nearly 50% in the past two decades, from 44% during 1995-2000 to 28% during 2001-2004. The decline in mortality has been attributed to more prompt and accurate diagnostic approaches, such as non-culture-based serum biomarkers and computed tomography (CT)-guided biopsy, as well as early initiation of novel broad-spectrum and less-toxic antifungal agents [Auberger et al. 2008].

Spectrum of fungal pathogens

Although aspergillosis and candidemia still account for the vast majority of cases of IFI among patients with AL, the spectrum of infectious fungal pathogens has changed dramatically in the last few decades. These changes include a reduction in the frequency of invasive candidiasis and increases in the frequency of fluconazole-resistant non-Candida albicans species (e.g. C. krusei and C. glabrata), non-Aspergillus fumigatus species, and non-Aspergillus species (e.g. Fusarium and Zygomycetes). The changes may be due to the increasing use of antimicrobial prophylaxis (including azole antifungal agents), autologous peripheral-blood stem cell transplants, and novel immunosuppressants [Singh, 2001].

For example, in a large retrospective study among patients in Italy with hematologic malignancies, two thirds of the 538 IFI cases were attributable to molds and the rest to yeasts [Pagano et al. 2006]. The majority of these mold infections were caused by Aspergillus species (e.g. A. fumigatus, A. flavus) (90%). Zygomycetes and Fusarium species each accounted for 4% of cases, and the remaining mold infections were due to Scedosporium, Acremonium, Penicillium, and Cladosporium species. Most of the yeast infections were caused by Candida species (91%); Candida nonalbicans species were responsible for over half the episodes of candidemia (57%). The remaining yeast infections were caused by Cryptococcus (4%), Trichosporon (4%), Rhodotorula and Hansenula (1%) species. The most common infectious species were not necessarily the most dangerous: the highest IFI-attributable mortality rate was associated with zygomycosis (64%), which was followed by fusariosis (53%), aspergillosis (42%), and candidemia (33%) [Pagano et al. 2006].

Site of infection

IFI can involve single organ or it can become disseminated. Potentially, IFI can involve most of the visceral organs in the body; however, infection of blood stream, lungs and sinuses is more common. Invasive candidiasis (IC) most frequently affects the blood stream [Hammond et al. 2010; Betts et al. 2006]. In a study of 27 neutropenic patients who had IC, the blood stream was the site of infection in 93% [Betts et al. 2006]. In contrast, invasive aspergillosis (IA) is most commonly found in the lungs and sinuses [Hammond et al. 2010; Pagano et al. 2010; Betts et al. 2006]. In a study of 41 neutropenic patients who had IA, the lungs (73%) were the most common site of infection, followed by the sinuses (17%) and disseminated disease (7%) [Betts et al. 2006]. The most frequent sites of infection for zygomycosis are the lungs and orbito-sinus-facial structures [Ruping et al. 2010; Pagano et al. 2004]. In a study of patients with hematologic malignancies, the most frequent sites of mucormycosis were the lungs (64%) and the orbito-sinus-facial structures (24%), while cerebral involvement and disseminated infection were observed in only 19% and 8% of the cases, respectively [Pagano et al. 2004].

Risk stratification

Among patients with AL, acute myeloid leukemia might impose a greater risk than acute lymphoid leukemia in the acquisition of IFI [Auberger et al. 2008; Chamilos et al. 2006; Pagano et al. 2006; Denning et al. 1998]. In a retrospective cohort study of 11,802 patients with hematologic malignancies, the incidence of IFI in patients with acute myeloid leukemia and acute lymphoid leukemia was 12% and 6.5%, respectively [Pagano et al. 2006]. The higher rate in acute myeloid leukemia could be related to a unique intrinsic functional defect or to a relative reduction in the absolute numbers of neutrophils at the start of treatment [Prentice et al. 2000].

Prentice and colleagues developed a risk stratification model for the acquisition of IFI (Table 1) [Prentice et al. 2000]. This model incorporates risk factors that were validated in a prospective study [McLintock et al. 2004] which had calculated event rates of approximately 22%, 10%, and 2% for the high, high-intermediate, and low-intermediate risk groups, respectively [McLintock et al. 2004].

Table 1.

Characteristics of groups at risk for invasive fungal infection.

Low-risk groups

Autologous HSCT

Childhood acute lymphoid leukemia (except for Pneumocystis carinii pneumonia)

Lymphoma

Intermediate-risk groups

Low-Intermediate

Moderate neutropenia (ANC 0.1–0.5 × 109/l) for < 3 weeks, lymphocyte count <0.5 × 109/l, and antibiotics (e.g. trimethoprim-sulfamethoxazole)

Older age

Central venous catheter

High-Intermediate

Colonization at >1 site or heavy colonization at 1 site

Moderate neutropenia (ANC <0.5 to >0.1 × 109/l) for >3 to <5 weeks

Acute myeloid leukemia

Total body irradiation

Allogeneic matched sibling HSCT

High-risk groups

Severe neutropenia (ANC) <0.1 × 109/l for >3 weeks

Colonization by Candida tropicalis

Allogeneic unrelated HSCT or mismatched donor HSCT

Graft-versus-host disease

Moderate neutropenia (ANC <0.5 × 109/l) for >5 weeks

Corticosteroids >1 mg/kg and mild neutropenia (ANC <1 × 109/l) for >1 week

Corticosteroids >2mg/kg for >2 weeks

High-dose cytarabine or fludarabine

ANC, absolute neutrophil count; HSCT, hematopoietic stem cell transplantation. Reproduced with permission from Prentice et al. [2000].

One of the main risk factors in the pathogenesis of IC is the colonization of yeast in the gastrointestinal mucosa with acute mucosal damage caused by cytotoxic drugs (e.g. high-dose cytarabine) [Prentice et al. 2000]. Other risk factors include: neutropenia, the use of broad-spectrum antibacterial therapy, bacteremia [Prentice et al. 2000], renal insufficiency, prolonged stay in an intensive care unit, receipt of total parenteral nutrition and a recent gastrointestinal surgical procedure [Pfaller et al. 2006]. Prior fungal exposure or colonization, prior exposure to antifungal therapy and the state of the underlying disease (for patients not in remission) are other important risk factors for IFI in general [Maertens, 2007].

Those at highest risk of acquiring IA are patients with AL and HSCT recipients [Patterson, 2009; Ruping et al. 2008; Muhlemann et al. 2005; Wiederhold et al. 2003]. For these groups, prolonged and profound neutropenia (neutrophil cell count >500/μl for >10 consecutive days), such as during the first induction cycle, is a major risk factor [Ruping et al. 2008; Muhlemann et al. 2005; Wiederhold et al. 2003]. Short intervals between neutropenic episodes also increases the risk of IA [Muhlemann et al. 2005], as do the use of immunosuppressants, including steroids [Patterson, 2009; Ruping et al. 2008; Wiederhold et al. 2003], graft-versus-host disease, cytomegalovirus infection [Wiederhold et al. 2003], hyperglycemia [Weiser et al. 2004], and high bone marrow iron stores [Kontoyiannis et al. 2007]. Compared with allogeneic HSCT, autologous HSCT is associated with a lower risk of IA because the period of pre-engraftment neutropenia is shorter and the possibility of graft-versus-host disease is absent [Wiederhold et al. 2003].

Polymorphisms in human genes that encode pattern recognition receptors (e.g., Dectin-1 genes) have been associated with a higher risk of Candida colonization [Marr, 2010]. Similarly, polymorphisms in the genes that encode Tolllike receptor 4, interleukin-10, and plasminogen have been shown to place an individual at higher risk of developing IA [Marr, 2010]. Better understanding of these immunogenetic factors will likely improve risk stratification.

Diagnosis

The European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group and the National Institute of Allergy and Infectious Diseases Mycoses Study Group (EORTC/MSG) has categorized the degree of certainty of an IFI diagnosis for patients with cancer and for recipients of HSCT as proven, probable, and possible.

This classification system is based on three elements: host factors, microbiological evidence, and clinical evidence. Although this categorization suggests the probability of IFI, it is intended for research rather than use in clinical practice [Ascioglu et al. 2002]. Because no one element is sufficient for diagnosing IFI, all three factors should be taken into consideration when infection is suspected.

Host factors

Five host factors greatly assist in stratifying a patient's risk level for IFI: (a) neutropenia (<500 neutrophils/mm³ for >10 days); (b) fever >96 hours refractory to broad-spectrum antibiotics; (c) temperature >38°Cor <36°C plus the presence of high-risk features, such as a history of proven or probable IFI during previous episodes of neutropenia or the coexistence of symptomatic acquired immunodeficiency syndrome (AIDS); (d) signs and symptoms indicating graft-versus-host disease, particularly severe (grade ≥2) or chronic extensive disease; and (e) prolonged (>3 weeks) use of corticosteroids in the previous 60 days [Ascioglu et al. 2002].

Microbiological evidence

The demonstration of a fungal element or positive culture from a site, in addition to associated pathologic, clinical, or radiologic evidence of tissue damage, establishes the diagnosis of an IFI [Ascioglu et al. 2002].

The presence of yeast in urine, mucous membrane, or sinuses might simply reflect the presence of contaminants or colonization. However, two positive results of urine culture for yeast or the detection of Candida casts in urine in the absence of a urinary catheter is considered significant for urinary tract IC [Ascioglu et al. 2002]. The presence of signs of IC categorizes IFI as proven, whereas in their absence IFI is instead considered probable [Ascioglu et al. 2002]. Regardless, candida is more likely to be isolated in blood culture in disseminated (58%) versus single-organ candidiasis (28%) [Berenguer et al. 1993]. The frequency of autopsy-proven candidiasis is not affected by whether the blood is collected from a central catheter or by peripheral venipuncture [Lecciones et al. 1992].

Blood culture has very poor sensitivity for detecting molds, e.g. IA. The presence of mold in urine or respiratory tract specimen could simply reflect the presence of contaminants or colonization.

However, a positive culture is considered significant for IA in the presence of temporally related clinical signs and symptoms compatible with the relevant organism. Results from histopathologic analysis can be diagnostic for IA and endemic dimorphic fungi [Ascioglu et al. 2002], with an overall prompt diagnosis and detailed evaluation of the surrounding tissue for microscopic features of tissue damage [Hayden et al. 2003]; however, without cultures and special stains histologic analysis may not allow differentiation between fungal species [Hayden et al. 2003; Kaufman et al. 1997].

Although microbiological studies are very useful in diagnosing IFI, the invasive procedures required to obtain tissue specimens from leukemia patients who have multiple comorbidities or coagulopathy can be challenging [Leventakos et al. 2010]. Furthermore, successful isolation of fungi can take a long time and is often limited by sampling error, the effects of recent antifungal therapy, and failure of the organisms to grow in culture. The latter may be due to the use of inappropriate media or methods for a given organism, poor handling of the clinical specimen prior to culture, or the inherent limitations of currently available microbiological techniques [Hayden et al. 2003].

Serologic evidence

Non-culture-based serum biomarkers, such as antigen assays, can greatly assist in the diagnosis of aspergillosis in a sample of bronchoalveolar lavage fluid (BAL), a sample of cerebrospinal, or ≥2 blood samples [Ascioglu et al. 2002]; cryptococcosis in a sample of blood [Ascioglu et al. 2002] or cerebrospinal fluid [Walsh and Chanock, 1998]; or histoplasmosis in a sample of blood, urine, or cerebrospinal fluid [Ascioglu et al. 2002]. However, false-positive reports are possible. For example, a positive cryptococcal antigen result may be secondary to infection by Trichosporon species [Walsh et al. 1993] or by Stomatococcus mucilaginosis [Chanock et al. 1993], or it may be secondary to the presence of circulating rheumatoid factor [Engler and Shea, 1994].

The colorimetric 1,3-β-D-glucan (BG) assay is another diagnostic tool. It has a sensitivity and specificity for IFI of 95% and 85%, respectively, with a cutoff value of 30 pg/ml. The sensitivity decreases and the specificity increases with the use of higher cutoff values [Obayashi et al. 2008].

In one study, the sensitivity was slightly lower and the specificity was slightly higher among patients with hematologic malignancies and HSCTrecipients than among inpatients with other high-risk host factors. However, using the results from the BG assay would have increased the certainty of the diagnosis from possible to probable in approximately half of the cases [Koo et al. 2009]. The usefulness of this test is limited by the lack of specificity in identifying any particular fungus and the production of low levels of BG by zygomycetes [Pickering et al. 2005]. False-positive results are possible with concomitant therapy using certain antimicrobials (e.g. some cephalosporins, carbapenems, and ampicillin-sulbactam) or with patients undergoing dialysis [Marty et al. 2006], because of cross interaction with Pseudomonas aeruginosa [Mennink-Kersten et al. 2008], or with prior use of antifungal therapy [Senn et al. 2008].

One of the most commonly used tests for detecting serologic biomarkers is the galactomannan (GM) assay, which is a double-sandwich enzyme-linked immunosorbent assay that uses a monoclonal antibody [Verweij et al. 1998] to detect low levels of GM [Maertens et al. 2001]. In one study, serial GM monitoring was found to have a sensitivity and specificity for IA of 92% and 95%, respectively, among patients with hematological disorders at high risk for IA [Maertens et al. 1999]. However, in a meta-analysis, GM assay was found to have a sensitivity and specificity of 70% and 92%, respectively, for IA among patients with hematological malignancy [Pfeiffer et al. 2006]. Both false-positive results (e.g. due to the use of certain β-lactam antibiotics) and false-negative results (e.g. due to the use of mold-active antifungal prophylaxis and corticosteroids) are possible [Marr et al. 2005, 2004; Mennink-Kersten et al. 2004]. Furthermore, GM assay in BAL specimen has shown promising results and adds to the sensitivity of the serum GM assay [Nguyen et al. 2010; Meersseman et al. 2008].

Combining the BG and GM assays is very useful for identifying false-positive results from each test and improves the specificity and positive predictive value for the diagnosis of IA in patients with hematologic malignancies who are at high risk of acquiring IFI [Pazos et al. 2005]. Furthermore, many promising molecular diagnostic studies for detecting IFI early are in development, such as in situ hybridization directed against ribosomal RNA sequences [Lass-Florl et al. 2007; Hayden et al. 2003; Torres et al. 2003] and polymerase chain reaction assays targeting gene regions [Einsele and Loeffler, 2008; McMullan et al. 2008; Donnelly, 2006; Lass-Florl et al. 2001].

Radiologic evidence

Small peripheral, target-like abscesses (bull'seye lesions) in the liver or spleen as demonstrated by CT, magnetic resonance imaging, or ultrasonography may assist in the diagnosis of chronic disseminated candidiasis [Ascioglu et al. 2002]. CT findings such as a halo sign, air crescent sign, or cavity within an area of consolidation in the absence of other cavitating infectious processes (Mycobacterium, Legionella, and Nocardia species) may help establish the diagnosis of an angio-invasive fungal pneumonia [Ascioglu et al. 2002]. A halo sign is an early feature of IFI that disappears within a week in approximately 75% of cases. In contrast, the air crescent sign appears about 2 weeks after an infection has been acquired in about two thirds of patients [Caillot et al. 2001]. Early use of high-resolution radiologic modalities to detect IFI in the early stage influences the outcome. In addition, the use of anti-fungal drugs when the halo sign is still present leads to improved outcome [Greene et al. 2007; Verweij et al. 2007; Caillot et al. 2001, 1997].

Prediction

The Candida score [Leon et al. 2006] and the clinical prediction rule [Ostrosky-Zeichner et al. 2007] are scoring systems that were devised in an attempt to identify critically ill patients at high risk of acquiring IC and to guide the use of empirical antifungal therapy. Their value in guiding such therapy for leukemia patients at high risk of IC (e.g. those who are neutropenic, immuno-suppressed, or receiving antifungal prophylaxis) is however, not clear. Studies of cancer patients have shown a higher frequency of non-C. albicans species among those with hematologic malignancies [Hachem et al. 2008a; Bodey et al. 2002; Kontoyiannis et al. 2001] or those receiving fluconazole prophylaxis [Bodey et al. 2002].

Among patients with leukemia, the presence of neutropenia and nodular and cavitary lesions on chest CT images are very helpful clues to the diagnosis of pulmonary IA [Hachem et al. 2006]. The combination of clinical features and radiologic findings may help in differentiating fungal infections. Prior use of voriconazole prophylaxis, concomitant sinusitis, multiple (>10) lung nodules, pleural effusion [Chamilos et al. 2005], or a ‘reverse halo sign’ [Wahba et al. 2008] on chest CT images favors the diagnosis of pulmonary zygomycosis over pulmonary IA.

Management

Early initiation of antifungal therapy is crucial for AL patients because delay is associated with increased mortality due to candidemia [Taur et al. 2010; Garey et al. 2006; Morrell et al. 2005] and due to invasive mold infection [Chamilos et al. 2008; von Eiff et al. 1995]. For cancer patients with candidemia, the incubation length i.e. the time to a positive culture (median, 32.1 hours) is an independent risk factor for inhospital mortality, which increases an estimated 1.025-fold for every additional hour [Taur et al. 2010]. The early use of diagnostic tests in the face of possible fungal infection can contribute to the early identification of IFI and initiation of antifungal therapy.

Empirical antifungal therapy

Broad-spectrum antifungal therapy is the standard of care for neutropenic patients who have fever lasting ≥5 days and are refractory to broad-spectrum antibiotics (i.e. are at high risk of acquiring IFI) [Hughes et al. 2002]. However, a wide range of inadequately treated nonfungal infections and noninfectious processes can precipitate fever, including drugs (e.g. chemotherapeutic agents), transfusion of blood products, thrombophlebitis [O'Grady et al. 1998], leukemia [Nakase et al. 1990], and graft-versus-host disease [Kim et al. 2004]). Conversely, a substantial proportion of patients with IFI, including the elderly and patients receiving antipyretic agents or anti-inflammatory drugs, can present with other signs of infection and no fever [O'Grady et al. 1998]. Although a thorough search for other possible causes of fever is always warranted, the threshold for initiating antifungal therapy should be low for patients otherwise at high risk for IFI.

The decision to initiate treatment and the choice of antifungal therapy should be based on risk stratification, local fungal epidemiologic patterns, concomitant comorbidities, drug-drug interactions, prior history of antifungal use, overall cost, and the pharmacologic profile of the anti-fungal armamentarium (Tables 2 and 3). When used empirically, voriconazole, caspofungin, and liposomal amphotericin B (AMB) seem to be equivalent in terms of efficacy and adverse events, whereas AMB deoxycholate is no longer recommended because of its toxicity and efficacy profiles [Goldberg et al. 2008; Klastersky, 2004].

Table 2.

Commonly used antifungal agents.

Agent	Mechanism of action [Groll et al. 1998]	Indication for therapy [Leventakos et al. 2010]	Dose and route [Leventakos et al. 2010]	Major adverse events	Comments
Azoles	Blocks the demethylation of lanosterol, thereby inhibiting ergosterol synthesis.			Well tolerated and safe; severe hepatotoxicity is uncommon [Dodds Ashley et al. 2006].	Inhibition of enzymes of cytochrome P450 system, particularly with voriconazole, results in many drug interactions. Oral absorption is influenced by food intake and gastric pH except with fluconazole [Dodds Ashley et al. 2006]. Owing to significant interpatient pharmacokinetic variability, the IDSA recommends plasma drug level monitoring of itraconazole, voriconazole, and posaconazole to optimize response and prevent toxicity [Walsh et al. 2008].
Voriconazole		1st line: aspergillosis, Scedosporium apiospermum 2nd line: Fusarium species	6 mg/kg q12 hr on days 1 and 2, then 4 mg/kg q12 hr; IV/PO	Reversible encephalopathy [Pascual et al. 2008], photopsia [Purkins et al. 2003; Herbrecht et al. 2002], cutaneous phototoxicity [Denning and Griffiths, 2001].	Should be administered on an empty stomach. Intravenous voriconazole should not be used in renal failure due to the accumulation of cyclodextrin [Dodds Ashley et al. 2006].
Posaconazole		2nd line: aspergillosis, zygomycosis, Fusarium species, Scedosporium apiospermum	800 mg/day in 2–4 divided doses; PO		Should be taken with a high-fat meal [Dodds Ashley et al. 2006]. Posaconazole is the only agent besides AMB that is active against zygomycetes [Leventakos et al. 2010; Dodds Ashley et al. 2006].
Itraconazole		2nd line: aspergillosis	200 mg q12 hr on days 1 and 2, then q24 hr; IV or PO	Nausea, diarrhea [Vandewoude et al. 1997], hypertension, hypokalemia, edema [Sharkey et al. 1991], congestive heart failure [Ahmad et al. 2001], hepatotoxicity [Dodds Ashley et al. 2006].	Intravenous itraconazole should not be used in renal failure due to the accumulation of cyclodextrin [Dodds Ashley et al. 2006].
Fluconazole		2nd line: candidiasisa	6-12 mg/kg/day; IV or PO	Reversible alopecia [Pappas et al. 1995].
Amphotericin B	Binds to ergosterol in the fungal cell membrane and disrupts cell permeability, resulting in rapid death.			Acute: infusion reaction (nausea, vomiting, rigors, fever, hypertension or hypotension, hypoxia, hyperkalemia) [Laniado-Laborin and Cabrales-Vargas, 2009; Dodds Ashley et al. 2006]; triad of symptom complexes: (a) chest spain, dyspnea, hypoxia; (b) abdominal, flank, or leg pain; (c) flushing and urticaria with lipid formulation [Roden et al. 2003].	Additive nephrotoxicity with use of other nephrotoxic agents (e.g. cyclosporine and aminoglycosides) [Dodds Ashley et al. 2006].
				Chronic: nephrotoxicity [Laniado-Laborin and Cabrales-Vargas, 2009; Dodds Ashley et al. 2006], hepatotoxicity [Dodds Ashley et al. 2006].
Deoxycholate		2nd line: candidiasis,a aspergillosis, zygomycosis	0.5-1 mg/kg/day; IV
Lipid formulation		1st line: aspergillosis, zygomycosis [Spellberg et al. 2009] 2nd line: candidiasisa	3-5 mg/kg/day; IV		Overall, lipid formulation has decreased nephrotoxicity compared with deoxycholate [Laniado-Laborin and Cabrales-Vargas, 2009]. Liposomal AMB appears to be less nephrotoxic compared to AMB lipid complex [Hachem et al. 2008b].
					Liposomal AMB has a higher tissue concentration in the central nervous system [Groll et al. 2000], whereas AMB lipid complex has a higher tissue concentration in lungs [Lewis et al. 2007; Groll et al. 2006], with potential enhanced fungicidal activity.
Echinocandins	Inhibits the production of (1→3)-β-D-glucan, an essential component of the fungal cell wall.	1st line: candidiasisa 2nd line: aspergillosis		Very well tolerated. Histamine-related infusion reaction is uncommon [Dodds Ashley et al. 2006].	Narrow spectrum. Very few drug interactions [Dodds Ashley et al. 2006]. Enhances immune response toward fungi, especially Zygomycetes species [Lamaris et al. 2008; Wheeler and Fink, 2006].
Caspofungin			70 mg on day 1, then 50 mg/day; IV		Undergoes hepatic degradation; requires dose modification in moderate-severe liver dysfunction [Dodds Ashley et al. 2006].
Micafungin			100-150 mg/day; IV		Undergoes hepatic degradation [Dodds Ashley et al. 2006].
Anidulafungin			200 mg on day 1, then 100 mg/day; IV		Undergoes slow nonenzymatic degradation in blood [Vazquez and Sobel, 2006].

IV, intravenous; PO, per os (by mouth].

^aThe choice of therapy in candidiasis is guided by several factors, such as the severity of illness, prior azole use, and need for mold coverage.

Table 3.

In vitro activity of antifungal agents.

	Agent
Species	Fluconazole	Itraconazole	Posaconazole	Voriconazole	Amphotericin B	Echino-candinsa
Aspergillus
A. fumigatus	−	+	+	+	+	+
A. flavus	−	+	+	+	+	+
A. terreus	−	+/−	+	+	+/−	+
Candida
C. albicans	+	+	+	+	+	+
C. krusei	−	+/−	+	+	+/−	+
C. glabrata	+/−	+/−	+/−	+/−	+	+
Otherb	+	+	+	+	+	+/−
Cryptococcus neoformans	+	+	+	+	+	−
Coccidioides species	+	+	+	+	+	−
Blastomyces dermatitidis	+/−	+	+	+	+	−
Histoplasma capsulatum	+	+	+	+	+	−
Fusarium species	−	−	+/−	+/−	+/−	−
Zygomycetes	−	+/−	+	−	+	−
Scedosporium
S. apiospermum	−	−	+	+	+/−	−
S. prolificans	−	−	−	−	−	−
Trichosporon	−	−	ND	+	+/−	−

+, In vitro activity; −, no in vitro activity; +/−, modest in vitro activity; ND, no data.

^aCaspofungin, micafungin, and anidulafungin.

^bCandida parapsilosis is less susceptible in vitro to echinocandins.

Reproduced with permission from Leventakos et al. [2010].

The spectrum of available antifungal agents should always be taken into consideration. Zygomycosis breakthrough infection among patients on voriconazole prophylaxis is being reported more frequently [Kontoyiannis et al. 2005; Imhof et al. 2004; Marty et al. 2004; Siwek et al. 2004]. In a recent study, individuals at high risk of IFI who received empirical anti-fungal coverage against zygomycetes had a superior survival rate compared with those who did not receive this therapy [Ruping et al. 2010]. Therefore, high-risk patients who receive voriconazole prophylaxis should also be considered for empirical antifungal therapy with coverage against zygomycosis.

Preemptive treatment

Both the early use of diagnostic tests in the face of possible fungal infection and the use of radiologic imaging, tissue sampling, and cytologic or histopathologic examination early in the disease process can assist in the early initiation of therapy and create an overall better outcome for the patient. This approach is called preemptive therapy. A prospective study of patients who had hematologic malignancies or solid-organ transplantation and were at high risk of IFI showed that calcofluor white staining, Aspergillus polymerase chain reaction, and GM enzyme-linked immunosorbent assay on CT-guided lung biopsy specimens can provide rapid and reliable information for the diagnosis of IFI [Lass-Florl et al. 2007]. This diagnostic intervention was also useful in confirming non-IFI diagnoses, differentiating between septated and nonseptated hyphae, and guiding therapy for infection with Aspergillus versus zygomycetes. In that study, at the time of diagnosis 84% of the patients with proven zygomycosis were not receiving drugs active against zygomycetes [Lass-Florl et al. 2007]. In another study, patients with hematologic malignancies for whom AMB-based therapy against zygomycosis had been delayed had higher mortality at 12 weeks than patients who had received treatment early [Chamilos et al. 2008].

Several studies have attempted to establish the feasibility of initiating preemptive therapy based on clinical symptoms, chest CT images, bronchoalveolar lavage assessment, and GM test results [Cordonnier et al. 2009; Segal et al. 2007; Maertens et al. 2005]. A recent multicenter, randomized, noninferiority trial of preemptive versus empirical antifungal therapy for febrile, neutropenic patients with hematologic malignancies who were at high risk of IFI showed no significant difference in mortality, with one exception: empirical treatment may have provided for better survival rates for patients who were receiving induction chemotherapy for acute myeloid leukemia [Cordonnier et al. 2009]. The major disadvantage of preemptive therapy is the possibility of missing cases of IFI [Cordonnier and Schwarzinger, 2009; Marr et al. 2009; Maertens et al. 2005]. Furthermore, delays in diagnostic interventions are more likely to occur outside of clinical trials.

Preemptive treatment of invasive candidiasis

The Infectious Diseases Society of America (IDSA) recommends the use of an echinocandin or a lipid formulation of AMB (LFAMB) as the primary treatment modality for candidemia in neutropenic patients [Pappas et al 2009]. Fluconazole is recommended for patients without recent exposure to azole and who are not critically ill. Voriconazole is recommended when additional coverage for molds is desired. More specifically, the IDSA recommends the use of echinocandin (rather than LFAMB) for infections due to C. glabrata; fluconazole or LFAMB for those caused by Candida parapsilosis; and echinocandin, LFAMB, or voriconazole for C. krusei. For candidemia without persistent fungemia or metastatic complications, therapy should continue for a total of 2 weeks after documented clearance of the fungus from the blood stream and after resolution of neutropenia and symptoms attributable to candidemia. For all cases of IC, removal of intravascular catheters is advised, albeit controversial [Pappas et al. 2009].

Preemptive treatment of invasive aspergillosis

The IDSA recommends voriconazole as the primary treatment modality for pulmonary IA or aspergillosis of the central nervous system; liposomal AMB may be used as an alternative primary therapy [Walsh et al 2008]. For salvage therapy, suggested agents include LFAMB, caspofungin, micafungin, posaconazole, and itraconazole. Although commonly used, combination therapy is not routinely recommended because clinical data are lacking on the efficacy and safety of combining antifungal agents for primary therapy; the decision to add another antifungal agent or to switch to another drug class for salvage therapy should be individualized. Surgical resection of Aspergillus-infected tissue may be useful for patients with lesions contiguous with the great vessels or pericardium, lesions causing hemoptysis from a single focus, or lesions causing erosion into the pleural space or ribs. Infected vascular catheters and prosthetic devices should be removed. Although the duration of treatment is not well defined, therapy should continue for a minimum of 6-12 weeks; for immunosuppressed patients, it should be continued throughout the period of immunosuppression and until lesions have resolved. For patients who had been successfully treated for IA and require subsequent immunosuppression, resumption of antifungal therapy at the time of the immunosuppressive treatment can prevent recurrent infection [Walsh et al. 2008].

Preemptive treatment of zygomycosis.

Zygomycosis has a very high mortality rate, especially among patients with hematologic malignancies and HSCT recipients [Spellberg et al. 2009]. In a review of 929 reported cases of zygomycosis, the mortality rate of patients who received AMB therapy was 39%, whereas 97% of untreated patients died [Roden et al. 2005]. These study results underline the importance of managing zygomycosis aggressively. Successful management depends on the reversal of underlying predisposing risk factors, prompt antifungal therapy, and surgical debridement where applicable [Spellberg and Ibrahim, 2010; Spellberg et al. 2009; Walsh and Kontoyiannis, 2008]. Surgery is an independent predictor of favorable outcome among patients with zygomycosis, and it might also improve outcome for patients receiving anti-fungal therapy [Singh et al. 2009; Roden et al. 2005; Lee et al. 1999].

AMB-based regimen is considered the cornerstone of primary therapy against zygomycosis [Spellberg and Ibrahim, 2010; Spellberg et al. 2009; Roden et al. 2005]. Although posaconazole is not suitable as the first-line agent [Spellberg and Ibrahim, 2010], as a salvage regimen, it has shown to have a response rate of 60% or even higher [Greenberg et al. 2006; van Burik et al. 2006]. In a retrospective study of non-neutropenic patients with rhino-orbital and cerebral mucormycosis, a combination polyenecaspofungin therapy had superior success (100% vs. 45%) and improved 4-month survival (100% vs. < 40%) compared with polyene monotherapy [Reed et al. 2008]. The role of the potential immunomodulatory beneficial effects of this combination therapy in neutropenic patients or immunosuppressed hosts is not clear [Walsh and Kontoyiannis, 2008]. Deferasirox, recombinant cytokines (e.g., granulocyte colony-stimulating factor, granulocyte-macrophage colony-stimulating factor, or interferon gamma); and hyperbaric oxygen might have some adjunctive value [Spellberg and Ibrahim, 2010], but their role in treating zygomycosis is not well established. The avoidance of iron supplementation and blood transfusion, if possible, is also advisable [Spellberg and Ibrahim, 2010].

Recent advances in antifungal therapy.

The limited efficacy of antifungal agents against certain IFIs in immunocompromised patients has spurred the development of new treatment strategies. These efforts include investigations into several immune therapies, the antifungal agent nikkomycin Z, and nanoparticle formulations of antifungal agents to enhance their solubility and biological availability and decrease their toxicity (e.g. AMB and itraconazole) [Ostrosky-Zeichner et al. 2010]. Other promising developments involve immunotherapy using radiolabeled antibodies (e.g., for Cryptococcus neoformans and Histoplamsa capsulatum) [Dadachova et al. 2004], vaccines (e.g., the recombinant N terminus of Als1p [rAls1p-N] vaccine against C. albicans) [Spellberg et al. 2005b], antibodies against fungal antigens (e.g. efungumab, a recombinant monoclonal antibody fragment to the heat shock protein of C. albicans) [Pachl et al. 2006], phagocytic cell-line-based immunotherapy (e.g. granulocyte transfusion or administration of activated HL-60 cells) [Lin et al. 2010; Anaissie, 2008; Spellberg et al. 2005a], and cytokines (e.g., granulocyte colony-stimulating factor, granulocyte-macrophage colony-stimulating factor, or interferon gamma) [Armstrong-James et al. 2010; van de Veerdonk et al. 2010; Anaissie, 2008].

Several studies have established the safety and adjunctive role of granulocyte transfusion in neutropenic patients with fungal infections [Al-Tanbal et al. 2010; Ang and Linn, 2010; Mousset et al. 2005]. A prospective, nonrandomized study of granulocyte transfusion in neutropenic patients with hematologic malignancies demonstrated control of acute life-threatening fungal infections at 30 days in 78% of episodes of fungal infection (24 of 31). The predominant fungal infections in that study were pulmonary aspergillosis and candidal sepsis [Mousset et al. 2005].

Prognosis

Uncontrolled leukemia is a poor prognostic factor for patients with IFI, and it may be important to minimize the time for which therapy of the hematologic malignancy is discontinued [Matt et al. 2004]. Achievement of remission of AL can also lead to the recovery of neutropenia and positively affect the outcome of fungal infection.

Prognosis of patients with invasive candidiasis

Predictors of unfavorable outcome for IC include old age (>60-65 years) [Hahn-Ast et al. 2010; Horn et al. 2010; Chen et al. 2006; Uzun and Anaissie, 2000], the presence and persistence of neutropenia [Horn et al. 2010; Uzun and Anaissie, 2000; Anaissie et al. 1998], multiorgan yeast dissemination [Uzun and Anaissie, 2000; Anaissie et al. 1998], corticosteroid therapy [Horn et al. 2010; Labelle et al. 2008; Chen et al. 2006], high Acute Physiology and Chronic Health Evaluation (APACHE) II or III scores [Horn et al. 2010; Labelle et al. 2008; Anaissie et al. 1998], and delayed treatment [Taur et al. 2010; Garey et al. 2006; Morrell et al. 2005]. Intensive care unit stay, sepsis at diagnosis, hyperalimentation, antimicrobial use [Chen et al. 2006], renal failure [Horn et al. 2010; Chen et al. 2006], and poor performance status [Uzun and Anaissie, 2000] are also associated with increased risk of death. In some studies removal of vascular catheters improved patient outcome [Horn et al. 2010; Labelle et al. 2008; Anaissie et al. 1998], but in another study removal of vascular access devices was not associated with better outcome except for a subset of patients with device-related candidemia [Chen et al. 2006]. Finally, C. glabrata infection was shown in one study to be an independent predictor of death [Chen et al. 2006], whereas in another recent study infection by Candida species did not influence patient survival or the rate of overall treatment success [Horn et al. 2010].

Prognosis of patients with invasive aspergillosis

For IA, poor prognostic factors include old age (>60 years) [Hahn-Ast et al. 2010], the progression or recurrence of underlying leukemia or cancer [Hahn-Ast et al. 2010; Nivoix et al. 2008; Matt et al. 2004], proven or probable (as opposed to possible) aspergillosis [Baddley et al. 2010; Hahn-Ast et al. 2010; Nivoix et al. 2008], steroid use, renal impairment, disseminated aspergillosis, diffuse pulmonary lesions, receipt of HSCT [Nivoix et al. 2008], and neutropenia [Baddley et al. 2010; Nivoix et al. 2008]. Surgical intervention (e.g. debridement or lung lobectomies) [Burgos et al. 2008] and early initiation of antifungal therapy [von Eiff et al. 1995] are independent predictors of survival.

Prognosis of patients with zygomycosis

Poor prognostic factors for zygomycosis include active malignancy, delayed AMB-based therapy (≥6 days after diagnosis), monocytopenia at the time of diagnosis [Chamilos et al. 2008], and renal failure and disseminated disease [Singh et al. 2009]. Use of liposomal AMB has been shown to improve outcome [Ruping et al. 2010; Singh et al. 2009]. Posaconazole-based salvage therapy [Chamilos et al. 2008], surgical intervention (e.g. debridement, or lung lobectomies) [Singh et al. 2009; Roden et al. 2005; Lee et al. 1999], and neutrophil recovery [Chamilos et al. 2008] all predict favorable outcome.

Conclusion

Trends in fungal infections are rapidly changing, and many novel diagnostic and therapeutic options are emerging. However, several areas in IFI management require further exploration: risk stratification based on immunogenetic factors, the role of new diagnostic modalities, prediction models for differentiating diverse fungal infections, and preemptive versus empirical therapy. Ongoing clinical trials are expected to provide further insight into the management of these complex and serious infections.