Treatment options of invasive fungal infections in adults

a Division of Infectious Diseases & Hospital Epidemiology, University Hospital of Basel, Basel; b Service of Infectious Diseases, Department of Internal Medicine, Centre Hospitalier Universitaire Vaudois (CHUV), Lausanne; c Institute of Microbiology, CHUV, Lausanne; d Service of Adult Intensive Medicine, CHUV, Lausanne; e Institute of Infectious Diseases, Inselspital, Berne; f Division of Infectious Diseases and Hospital Epidemiology, University Hospital of Zürich, Zurich; g Division of Infectious Diseases and h Infection Control Program, Department of Internal Medicine, University Hospital of Geneva, Geneva

A panel of infectious disease specialists, clinical microbiologists and hospital epidemiologists of the five Swiss university hospitals reviewed the current literature on the treatment of invasive fungal infections in adults and formulated guidelines for the management of patients in Switzerland.For empirical therapy of Candida bloodstream infection, fluconazole is the drug of choice in non-neutropenic patients with no severe sepsis or septic shock or recent exposure to azoles.Amphotericin B deoxycholate or caspofungin would be the treatment option for patients with previous azole exposure.In neutropenic patients, empirical therapy with amphotericin B deoxycholate is considered first choice.In patients with severe sepsis and septic shock, caspofungin is the drug of first choice.For therapy of microbiologically-documented Candida infection, fluconazole is the drug of choice for infections due to C. albicans, C. tropicalis or C. parapsilosis.When infections are caused by C. glabrata or by C. krusei, caspofungin or amphotericin B deoxycholate are first line therapies.Treatment guidelines for invasive aspergillosis (IA) were stratified into primary therapy, salvage therapy and combination therapy in critically ill patients.Voriconazole is recommended for primary (ie upfront) therapy.Caspofungin, voriconazole (if not used for primary therapy) or liposomal amphotericin B are recommended for salvage therapy for refractory disease.Combination therapy with caspofungin plus voriconazole or liposomal amphotericin B should be considered in critically ill patients.Amphotericin B deoxycholate is recommended as initial therapy for the empirical therapy in patients with neutropenia and persistent fever with close monitoring of adverse events.

Key words: guidelines; candida; invasive aspergillosis; antifungal agents
Candida and Aspergillus are the most common causes of invasive fungal infections, accounting for 70-90% and 10-20% of all invasive mycoses, respectively [1].In the US, the incidence of fungal sepsis has increased three-fold between 1979 and 2000 [2].Invasive candidiasis and aspergillosis are associated with substantial morbidity and high mortality (40-60% and 60-90%, respectively), prolonged hospital stay and increased health care costs [3][4][5][6][7].Early diagnosis and prompt initiation of antifungal therapy are thus essential to reduce morbidity and mortality.
For decades, amphotericin B deoxycholate, has been standard therapy for invasive fungal in-fections.Unfortunately, amphotericin B deoxycholate is often poorly tolerated and associated with infusion-related acute reactions and nephrotoxicity.In the late 1970s and 1980s, the emergence of azoles (first miconazole and ketoconazole and then fluconazole and itraconazole), a new class of antifungal agents inhibiting the synthesis of the cell membrane, provided an alternative therapeutic strategy to amphotericin B deoxycholate.In recent years, several new antifungal agents have become available offering additional therapeutic options for the management of invasive fungal infections.These include lipid formulations (colloidal dispersion, lipid-complex and liposomal) of Introduction amphotericin B, new azoles (voriconazole, and posaconazole) and echinocandins (caspofungin, micafungin and anidulafungin) [8][9][10][11][12].
A recent publication has shown the utility of using treatment guidelines to improve the outcome of patients with invasive fungal infections [13].Therefore the aim of the present article is to review the current literature on antifungal ther-apy and to formulate treatment recommendations for the management of the two most frequent invasive fungal infections occurring in surgical, critically ill or immunocompromised patients, namely invasive candidiasis and invasive as pergillosis and for empirical therapy of suspected fungal infections in patients with neutropenia and persistent fever.
The article was prepared by a group of infectious diseases specialists, clinical microbiologists and hospital epidemiologists from the five Swiss university hospitals with expertise in the diagnosis and the management of fungal infections.This group was founded in 2000 under the name of the Fungal Infection Network of Switzerland (FUNGINOS).The names of participating individuals and institutions of FUNGINOS can be found in the Appendix of the first article published by the group on the epidemiology of Candida infections [14].FUNGINOS was created with the overall objective to promote biomedical research on invasive fungal infections in Switzerland.The recommendations were presented and discussed at meetings of the FUNGINOS group and of the Swiss Society of Infectious Diseases.Modifications were circulated electronically and agreed upon as part of an iterative process until consensus was reached by a majority of Steering Committee members.Treatment recommendations were rated according to a standard scoring scheme to illustrate the strength of the supporting evidence and the quality of the underlying data as follows: Level of recommendation: A = good evidence; B = moderate evidence; C = poor evidence, Quality of recommendation: I = randomised, controlled, clinical study; II = non-randomised, controlled clinical study or cohort study or case-control study; III = case series or expert opinion [15].

Epidemiology
Candida species are the most frequent cause of fungal infections (70 to 90% of cases) and account for 5 to 15% of health-care associated infections [16,17].Risk factors for fungal sepsis include prematurity, low birth weight, disruption of cutaneous or mucosal barriers (such as surgical wounds, burns or chemotherapy-induced mucositis), indwelling intravascular or urinary catheters, defects of cellmediated immunity, broad-spectrum antibiotic therapy, parenteral nutrition, colonisation with Candida at multiple sites, surgery (especially gastrointestinal interventions) or haemodialysis [18][19][20][21].Critically ill and severely immunocompromised patients are at particularly high risk of invasive candidiasis.Candidaemia, one of the most frequent clinical manifestations of candidiasis, is associated with significant morbidity, prolonged hospital stay, high crude and attributable mortality (40% to 60% and 20 to 40%, respectively) and increased health care costs [3,6].In a multivariate analysis of mortality associated with bloodstream infections, Candida was found to be the sole pathogen independently associated with fatal outcome [22].A nationwide population surveillance study conducted in 50 hospitals in the US between 1995 and 2002 has confirmed that Candida is the fourth most common cause of bloodstream infections, representing 9.5% of all episodes [17].In contrast, recent data from Europe have shown that Candida species were a less common cause of bloodstream infections.Candida species ranked number seven accounting for 3% of all bloodstream infections in a 10-year retrospective study carried out in Switzerland by the Fungal Infection Network of Switzerland (FUNGINOS) [14].In this survey, one third of all episodes of candidaemia occurred in ICUs (intensive care unit), one third in medical wards, including haematology-oncology units, and one fourth in surgical wards [14].

Antifungal therapy
Rapid initiation of appropriate antifungal therapy is essential for the control of systemic Candida infections and has been shown to reduce mortality [41][42][43].Management guidelines have recommended that all patients with candidaemia, defined as the detection of Candida in at least one blood culture, should be treated with antifungal agents [20,44,45].

Polyenes
Amphotericin B deoxycholate, a fungicidal agent with broad-spectrum antifungal activity that acts on ergosterol of the fungal cell membrane, has been used for decades as the treatment of choice of invasive candidiasis.It is usually used at a dose ranging between 0.6 and 1 mg/kg/d i.v.Unfortunately, amphotericin B deoxycholate is often poorly tolerated being associated with infusion-re-lated acute reactions (eg chills, fever, hypoxaemia, and hypotension), especially when administered over a short period of time (ie 4 to 6 hours), and with nephrotoxicity (decreased glomerular filtration and tubular wasting of potassium, magnesium, and bicarbonate).However, four recent studies have shown that administration of amphotericin B deoxycholate as a continuous infusion over 24 hours with saline loading reduced infusionrelated reactions and renal impairment, including in allogenic stem cell transplant patients receiving cyclosporin A [46][47][48][49].Also, a multivariate analysis on the epidemiology of renal toxicity in 494 patients receiving amphotericin B deoxycholate showed that male gender, body weight 090 kg, chronic renal disease, treatment with an aminoglycosides or cyclosporin and doses of amphotericin B 035 mg/day were independent risk factors for nephrotoxiciy.Of note, the incidence of nephrotoxicity rose with the increasing number of risk factors suggesting that alternative therapy might be appropriate in patients with 2 or more risk factors [50,51].
Lipid formulations of amphotericin B (colloidal dispersion, lipid-complex and liposomal) are better tolerated than amphotericin B deoxycholate and have been used mainly in patients intolerant to conventional amphotericin B or unlikely to tolerate it because of already altered renal function [9,[52][53][54][55].Only liposomal amphotericin B (Am-Bisome ® ) is available in Switzerland.Few studies have compared the efficacy of amphotericin B deoxycholate with that of lipid formulations for the treatment of patients with invasive candidiasis.Small non-comparative studies suggest that lipid formulations of amphotericin B are as efficacious as conventional amphotericin B [56][57][58][59].High costs, a relative paucity of clinical data and the existence of alternative antifungal therapies (azoles and echinocandins) explain why lipid formulations have been generally used as second-line therapy in patients with refractory invasive candidiasis [20,45].

Azoles
In the late 1980s triazoles rapidly became standard therapy for invasive candidiasis.Azoles inhibit the synthesis of ergosterol of the fungal cell membrane.In vitro, these compounds are fungistatic against Candida species.Several clinical studies have compared the efficacy and safety of azoles General pattern of susceptibility of Candida species (adapted from [20]).
with that of amphotericin B deoxycholate for the treatment of candidaemia in non-neutropenic patients, but few data are available on the treatment of neutropenic patients.Fluconazole (Diflucan ® ) is available as intravenous and oral formulations.Oral fluconazole is very well absorbed and the bioavailability is not influenced by H 2 blocking agents.The daily dose should be reduced by half in patients with creatinine clearance of less than 50 ml/min and by 75% if less than 20 ml/min.A loading dose of twice the daily dose is recommended.In a multicentre study of 206 non-neutropenic patients with candidaemia, fluconazole was found to be as efficacious as (success rates: 72% vs 79%, respectively) and better tolerated than amphotericin B deoxycholate [60].Given the broad use of fluconazole, an increasing incidence of infections due to non-albicans Candida species with reduced, dose-dependent, susceptibility (C.glabrata) or intrinsic resistance (C.krusei) to azoles has been reported in the late 1990s [17,25,61].Data on the efficacy of high doses (800 to 1200 mg) of fluconazole for the treatment of suspected azole-resistant Candida bloodstream infections are lacking.
Itraconazole is available as oral and intravenous formulations.However, the poor bioavailability (enhanced by food) of its original oral formulation has limited its use and no comparative clinical studies have been performed in patients with candidaemia.An intravenous formulation of itraconazole has recently become available, but few clinical data have assessed its efficacy and safety.Given the variation in its bioavailability and potential for drug-drug interactions (eg with rifampicin, anticonvulsants, protease-inhibitors, macrolides) it is recommended to monitor itraconazole blood levels during therapy.
Voriconazole is available as oral (bioavailability 60 to 100%) and intravenous formulations [10].It is metabolised in the liver via the P-450 system (CYP2C9, CYP3A4 and CYP2C19), hence the potential for multiple drug interactions (eg with rifampicin, anticonvulsants, sirolimus, tacrolimus, cyclosporin, oral anticoagulants, statins, omeprazole, protease-inhibitors, NNRTI).There is no need to alter voriconazole doses in patients with renal impairment receiving the oral formulation.However, administration of intravenous voriconazole in patients with moderate renal insufficiency (creatinine clearance of less than 50 ml/min) may result in accumulation of the b-cyclodextrin used to solubilise voriconazole.Adverse events include transient and fully reversible, non-sightthreatening visual disturbances (in about 30 to 40% of patients), rash, hepatitis, and hallucinations.Voriconazole exhibits excellent in vitro and in vivo activities against Candida species [39].Clinical data from immunocompromised (mainly HIV-positive) patients with oropharyngeal and/or oesophageal candidiasis suggest that voriconazole has excellent activity against fluconazole-susceptible and fluconazole-resistant Candida infections: 98% success of voriconazole versus 95% success of fluconazole [62].Voriconazole salvage therapy showed an overall success rate of 55% in patients with refractory systemic candidiasis [63].In a randomised, open-label, comparative multicentre non-inferiority trial in 422 patients with invasive Candida infections (of whom >95% had candidaemia), voriconazole was shown to be at least as effective as a regimen of amphotericin B deoxycholate followed by intravenous or oral fluconazole [64].Of note, for C. tropicalis infections, success rates were higher with voriconazole than with amphotericin B/fluconazole (P = 0.03).
Posaconazole is a new azole with broad-spectrum antifungal activity against Candida, Aspergillus, and other emerging molds including Fusarium and the zygomycetes.The hepatic metabolic pathways of posaconazole differ from those of other azoles: glucuronation plays a major role, while enzymes of the CYP450 system are of secondary importance, which may decrease the risk of clinically significant drug-drug interactions [65].The drug is available in oral form with variable bioavailability, which can be significantly improved (ie up to 90%) by food intake [66].In 199 HIV-positive patients with azole-refractory oropharyngeal or oesophageal candidiasis, posaconazole 400 mg twice daily was clinically successful in 75% of cases (modified ITT analysis) and 81.6% of cases (evaluable population) [67].In another clinical study, the success rate of posaconazole as salvage therapy in patients with invasive candidiasis was 47.8% [68].In a long-term safety study in 102 patients treated with posaconazole for salvage therapy of invasive mycoses serious adverse events occurred in 12/102 (12%) patients, but only in one case posaconazole therapy was discontinued [69].Although posaconazole will be soon approved for salvage therapy of refractory mycoses, clinical data from comparative studies on first-line therapy of candidiasis are needed.

Echinocandins
Echinocandins are a new class of parenteral antifungal agents that act by inhibiting the glucane synthase, the enzyme responsible for the synthesis of b-(1,3)-D-glucan in the fungal cell wall [12].These compounds are fungicidal in vitro against C. albicans and non-albicans Candida species.No cross-resistance with azoles has been reported.
Caspofungin is the first echinocandin licensed for the treatment of invasive mycoses, including candidiasis [12].It is only available as an intravenous formulation.Caspofungin is taken up by the liver and is slowly degraded by hydrolysis and N-acetylation.Therefore, there is no need for dose adjustments in patients with renal insufficiency, but dose reduction to 35 mg is recommended in patients with moderate hepatic dysfunction.No data are available in patients with severe impairment of liver function or in children.Safety profile of caspofungin is excellent with few reported adverse events (mostly abnormal liver function tests, phlebitis or histamine-like reactions).Drug-drug interactions with some medications have been observed (eg with rifampicin, anticonvulsants, tacrolimus, cyclosporin, proteaseinhibitors, NNRTI).In immunocompromised (mainly HIV-positive) patients with oropharyngeal and/or oesophageal candidiasis caspofungin was found to be as effective (success rates: 70 to 90%) as amphotericin B deoxycholate (success rate: 65%) or fluconazole (success rate: 80%) [70][71][72].In a large multicentre trial that included 239 patients (of whom 24 were neutropenic) with invasive candidiasis, caspofungin was at least as efficacious as and less toxic than amphotericin B deoxycholate (success rates: 73% vs 62%; discontinuation for adverse events: 3% vs 23%) [73].Of note, the success rates of caspofungin against C. glabrata and C. krusei were comparable to those obtained in azole-susceptible Candida species.
Micafungin and anidulafungin.Micafungin has been approved in Japan and recently by the FDA for prophylaxis of Candida infections in patients undergoing haematopoietic stem cell transplantation and for the treatment of oesophageal candidiasis.Approval in Europe is pending.Anidulafungin is in clinical development.In randomised, double-blind studies, the efficacy of micafungin (50, 100 or 150 mg/d; success rates: 70-90%, or anidulafungin (100 mg, followed by 50 mg/d; success rate: 97%) was found to be comparable to that of fluconazole (200 mg, followed by 100 mg/d; success 85-95%) for the treatment of oesophageal candidiasis in immunocompromised (mainly HIV-positive) patients [74,75].Micafungin was reported to yield 100% response rates in a small non-comparative series of six cases of candidaemia [76].In an open-label, non-comparative, international study of patients with newly diagnosed or refractory candidiasis, micafugin was associated with an overall success rate of 83.3% [77].Success rates of 84 to 90% have been observed in a phase II, dose-finding, study of anidulafungin for the treatment of 68 patients with invasive candidiasis [78].In a randomised, double-blind phase III study of 245 adult patients with invasive candidiasis anidulafungin treatment was statistically superior to fluconazole (success rates at end of i.v.therapy 75.6% vs 60.2%, difference 15.4%, 95% Cl 3.8 to 27).Persistence of candidaemia at the end of i.v.therapy was reported in 6.3% and 14.4% of patients receiving anidulafungin and fluconazole, respectively.Survival at 6 weeks was higher in the anidu-lafungin than in the fluconazole treatment group (75% vs 69%) [79].

Combinations of antifungal agents
Given the poor prognosis of Candida sepsis, clinicians have shown great interest for using combinations of antifungal agents of different classes for the treatment of critically ill patients.However, until now there have been relatively few studies on combinations of antifungal agents in invasive candidiasis.Some experts recommend the combination of amphotericin B deoxycholate and 5-flucytosine for the treatment of life-threatening Candida infections [20,44,45].This recommendation is based on the demonstration of synergistic effects in in vitro studies and in experimental animal models of systemic candidiasis and Candida endocarditis [80][81][82].Combinations of amphotericin B and 5-flucytosine have been shown to be superior to monotherapy for therapy with cryptococcal infections in immunocompromised patients, but there are few data in patients with invasive Candida infections [83][84][85][86][87][88][89][90].Given the lack of evidence on clinical efficacy, concerns about toxicity (impairment of renal function by amphotericin B may lead to accumulation of flucytosine and increased toxicity) and the recent development of new therapeutic options in critically patients, combined amphotericin B deoxycholate and 5-flucytosine therapy is not recommended.Despite in vivo antagonism between amphotericin B and azoles in experimental models of aspergillosis [91][92][93], data obtained in animal models of invasive candidiasis and Candida endocarditis did not show antagonism, but also no synergism [91][92][93].In a randomised, double-blind study in 219 non-neutropenic patients with candidaemia, fluconazole (800 mg/d i.v.) was compared to a combination of fluconazole (800 mg/d i.v.) and amphotericin B deoxycholate (0.7 mg/kg/d i.v.) [94].At first glance, the efficacy of combination therapy was slightly superior to that of monotherapy (success 69% vs 56%).However, there were statistically significant differences in baseline covariates between the two treatment groups, such as diseases severity as measured by the APACHE II score, which was lower in the combination group.While awaiting the results of prospective, randomised clinical trials demonstrating that combinations of antifungals are superior to and no more toxic than treatment with single agents, the indiscriminate use of combined regimens should be discouraged.

Empirical therapy for Candida bloodstream infections (table 2A and figure 1A)
An algorithm for choosing an empirical antifungal treatment for patients with Candida bloodstream infection (before information about species identification and susceptibility becomes available) is shown in figure 1A.Several factors should be taken into account that may help physicians when choosing an antifungal agent for empirical therapy.These includes local epidemiological data, prior exposure to antifungal therapy, presence of neutropenia, severity of the clinical pres-

Treatment recommendations
entation (presence of sepsis, severe sepsis or septic shock) and presence of underlying diseases or organ dysfunctions that may affect drug metabolism or increase the risk of drug-related toxicities.Treatment recommendations have been developed for four clinical conditions covering most of the circumstances encountered by practicing physicians: 1) absence of neutropenia, of severe sepsis or septic shock, or of recent exposure to azoles, 2) previous exposure to azoles, 3) presence of neutropenia, and 4) presence of severe sepsis or septic shock.
1 1 Few clinical data are available on the use of azoles and echinocandins in neutropenic patients with documented invasive candidiasis.In vitro, azoles are fungistatic, echinocandis are fungicidal.In some experimental models (eg Candida endocarditis, disseminated candidiasis in neutropenic animals), azoles are less efficacious than amphotericin B or echinocandins. 2 Amphotericin B deoxycholate (Fungizone ® ) not recommended in critically ill patients with severe sepsis/septic shock: risk of acute nephrotoxicity or of underdosing due to infusion-related toxicity.Caspofungin (Cancidas ® ) = first choice or alternative, respectively, in this setting. 3According to susceptibility testing.Some experts would add voriconazole to the list of first choice agents for the treatment of C. glabrata infections. 4For C. glabrata switch to oral itraconazole or voriconazole only after resolution of clinical symptoms/signs of infection and only if susceptibility test shows MIC ≤0.5 mg/L for itraconazole) or ≤1.0 mg/L for voriconazole, respectively.

Table 2B
Therapy for microbiologically-documented Candida infections.
ommended as first choice agents for empirical therapy.Contraindications for using amphotericin B deoxycholate are listed in the footnote of table 4. Alternative therapies are caspofungin or liposomal amphotericin B.
1.3.Neutropenia.Few data are available on the clinical efficacy of antifungal agents in neutropenic patients with candidaemia.Data from experimental models in animals with Candida endocarditis or invasive candidiasis in the context of neutropenia suggest that fungicidal agents, such as amphotericin B or caspofungin, might be more efficacious than fungistatic drugs like azoles [93,95].Moreover, lower reponse rates have been reported in neutropenic compared with non-neutropenic patients with candidaemia [25,96,97].Given that it might be preferable to use fungicidal agents in neutropenic patients, amphotericin B deoxycholate is proposed as first choice in patients without contraindications (see footnote table 4).Lipid formulations of amphotericin B and caspofungin are alternative therapies.
1.4.Severe sepsis and septic shock.Empirical antifungal therapy active against albicans and non-albicans Candida species (including species with decreased susceptibility to azoles accounting for 15% to 20% of all episodes in Switzerland) is required in patients with yeast bloodstream infections complicated by severe sepsis or septic shock [14].Nephrotoxic agents should be avoided given the risk of development of renal failure in the haemodynamically unstable, septic patient.Despite limited clinical experience in this setting, caspofungin is the drug of first choice due to its excellent efficacy and safety profiles.Alternative therapies comprise lipid formulations of amphotericin B or voriconazole in patients not recently exposed to azoles.

Therapy for microbiologically-documented Candida infections (table 2B and figure 1B)
Although antifungal susceptibility tests are not performed routinely in all medical centres, susceptibility can usually be estimated upon identification of the Candida species.
For  [98] (J.Bille, unpublished observation).Susceptibility should therefore be checked before using voriconazole for the treatment of C. glabrata infections.As soon as the patient is clinically stable and is able to take an oral medication, a switch to either itraconazole or voriconazole should be considered, provided that the isolate is susceptible.
Duration of therapy.For primary bloodstream Candida infections, it is recommended to treat patients for two weeks after the last positive blood culture.In case of invasive candidiasis with involvement of one or more organs (eg liver, spleen, lung, skin, eye, or bone), therapy should be continued until complete resolution of clinical symptoms and signs at all sites of infection.

Epidemiology
Invasive aspergillosis (IA) is the second most common invasive mycosis accounting for 10 to 20% of all cases.This potentially life-threatening infection occurs mainly in immmunocompromised hosts, including neutropenic cancer patients, haematopoetic stem cell (HSCT) or solid organ transplant recipients, especially those with severe graft-versushost disease (GVHD), patients treated with immunosuppressive therapies (corticosteroids, cytotoxic agents or monoclonal antibodies acting as anti-immune mediators) [7].The highest incidence (10 to 20%) and mortality (60 to 90%) rates of IA have been reported in allogeneic HSCT and in heart, lung or heart/lung transplant recipients [7,99,100].Apart from the classical immunocompromised host, IA has been recently recognised to occur in a growing number of critically ill patients receiving corticosteroids for treatment of chronic lung diseases, including mechanically-ventilated intensive care unit patients [101].In a recent report, 3.6% of 1850 medical ICU patients developed invasive aspergillosis, which was associated with a mortality of 80% [102].Of note, several of these patients did not have classical risk factors, but various underlying conditions including liver cirrhosis.The most frequent site of IA is the lower respiratory tract accounting for about 80 to 90% of all cases.Other classical clinical manifestations include sinusitis, central nervous system (CNS) infections and disseminated infection involving multiple sites [7,99].
Standardised definitions of IA were recently established by an expert panel of the EORTC and the MSG [103].Based on clinical, radiological and microbiological criteria, cases of IA are classified into one of three categories: possible, probable or proven IA.However, these definitions have been primarily developed for clinical research, and their usefulness for management strategies remains to be determined.

Microbiology
A. fumigatus is the predominant Aspergillus species isolated in patients with IA (accounting for more than 90% of cases in many series), followed by A. flavus, A. terreus and A. niger [104].However, a shift in the epidemiology of Aspergillus infections has been recently observed with an increasing frequency of infections caused by non-fumigatus Aspergillus species, such as A. terreus which is resistant to amphotericin B [105]. A. terreus is more often a health-care-associated than a communityacquired infection and has been associated with lower response rates to antifungal therapy when compared to A. fumigatus (28% versus 39%).Significant progress has been made in in vitro susceptibility testing for molds, including Aspergillus [106][107][108][109]. Yet, it is still a research tool performed in a few laboratories with special expertise in the field and clinical studies on in vitro and in vivo correlations remain to be conducted.

Antifungal therapy
Antifungal agents licensed for the treatment of invasive aspergillosis (IA) are amphotericin B deoxycholate, the lipid formulations of amphotericin B, voriconazole, itraconazole and caspofungin.Clinical data on their efficacy and safety profiles are mainly derived from patients with pulmonary aspergillosis, the most frequent form of IA.Since the mortality of IA ranges between 60 and 90%, an aggressive diagnostic approach (radiological imaging, cultures, determination of galactomannan and histopathological examinations of biological fluids or tissue biopsies), followed by prompt institution of an appropriate first-line antifungal therapy is critical [7,104,110,111].

Amphotericin B
For decades, amphotericin B deoxycholate has been standard therapy of IA [112], but response rates have been disappointing, especially in patients with profound and persistent neutropenia or with graft-versus-host disease.In a review of 1941 cases, crude mortality rates of 60%, 26%, and 88% were reported in immunocompromised patients with pulmonary, sinus, and cerebral aspergillosis, respectively.In patients with invasive pulmonary aspergillosis, who had been treated for more than 14 days, response rates were 83% in heart or renal transplant recipients, 54% in leukaemic patients, 33% in HSCT, and 20% in liver transplant recipients [7].Underdosing and treatment interruptions due to infusion-related and renal toxicity may affect the efficacy of amphotericin B deoxycholate therapy.In particular, nephrotoxicity requiring haemodialysis has been independently associated with death [113].
Lipid-based formulations of amphotericin B are better tolerated than amphotericin B deoxycholate.Outside clinical trials, lipid formulations of amphotericin B have been used primarily as salvage therapy in patients intolerant of or failing therapy with amphotericin B deoxycholate [52,114,115].Although lipid formulations of amphotericin B have recently become the new standard comparator in clinical trials of empirical therapy in persistently febrile neutropenia, there are very few randomised studies comparing lipid forms to amphotericin B deoxycholate as first-line therapy of IA.In a randomised double-blind study in 174 immunocompromised patients with invasive aspergillosis similar success rates were obtained with amphotericin B colloidal dispersion (35%) and amphotericin B deoxycholate (35%) [116].A study comparing a 1 to a 4 mg/kg/d i.v.dose of liposomal amphotericin B (AmBisome ® ) as first-line therapy of IA in 87 patients yielded similar survival rates (43 and 37%) [117].Lastly, a recent prospective, randomised study comparing a high dose (10 mg/kg/d) to a standard dose (3 mg/kg/d) of liposomal amphotericin B for the treatment of invasive filamentous fungal infections yielded similar success and survival rates [118].Overall, all the available data suggest that lipid formulations are at least as efficacious as and better tolerated than amphotericin B deoxycholate for the treatment of IA [53,112].

Azoles
Itraconazole has been used as first-line treatment of invasive aspergillosis in few small noncomparative studies.In 76 patients with IA, oral itraconazole therapy resulted in a response rate of 39% [119].A response rate of 48% was reported in a study of 31 patients with IA who were treated with itraconazole [120].The highly variable bioavailability of oral itraconazole and risk of drug-drug interactions (eg with immunosuppressive drugs) also are serious drawbacks for the treatment of patients with life-threatening infections.Monitoring of blood levels is therefore mandatory.Itraconazole has thus been mainly used as stepdown oral therapy in patients with IA who had responded to initial i.v.antifungal therapy [121].
Voriconazole is the present treatment of first choice for IA, and is available as intravenous and oral formulations.In a non-comparative study of first-or second-line treatment of IA in 116 patients, treatment with voriconazole was found to be associated with response rates of 48% [122].The efficacy and safety of upfront therapy with voriconazole was compared to that of amphotericin B deoxycholate in the largest (277 patients), prospective, multicentre, randomised, comparative study of IA ever performed.After 12 weeks of protocol therapy with or without switch to other licensed antifungal agents due to failure or toxicity, success and survival rates were significantly better in the voriconazole group than in the amphotericin B group (53% versus 32% and 71% versus 58%, respectively) [123].Moreover, fewer severe adverse events occurred in patients treated with voriconazole than in those treated with amphotericin B. There is substantial intraand inter-individual variability in voriconazole blood levels: low levels have been observed, in particular in children, whereas elevated levels have been measured in patients with adverse events [122,[124][125][126][127][128][129].Interactions with co-medications with hepatic metabolism (listed in the section on invasive candidiasis) may necessitate close clinical monitoring.
Posaconazole, a new azole antifungal agent was studied as salvage therapy of invasive aspergillosis in 107 patients refractory to or intolerant of first-line antifungal therapy [68].Posaconazole was successful in 42% of patients.Response rates in high-risk subgroups were 24% in neutropenic patients and 31% in allogeneic HSCT recipients.Overall the 12-month survival was 45%.A salvage, open-label, compassionate study in 23 solid-organ transplant recipients with invasive mycoses refractory to or intolerant of standard therapy reported a complete or partial response in 57% of cases [130].Drug-drug interactions increasing blood levels of cyclosporine or tacrolimus were observed in 3 cases and lead to discontinuation of posaconazole in one patient.A recent double-blind multicentre study in 600 allogeneic HSCT recipients compared posaconazole and fluconazole for prophylaxis of invasive mycoses [131].The incidence of invasive aspergillosis during the 16-week study period was significantly lower in the posaconazole group (2% vs 7%).The mortality due to invasive mycoses was 1% vs 4%, respectively.Adverse events occurred in 33% of cases in both groups, and resulted in therapy discontinua-Polyenes tion in 3% of patients receiving posaconazole vs 8% in those receiving fluconazole.Awaiting results from comparative studies of first-line therapy of invasive aspergillosis, posaconazole will be soon available as salvage therapy of refractory invasive mycosis.

Echinocandins
Caspofungin has been licensed for salvage therapy of IA after it was shown to be efficacious and safe in 83 patients with haematological malignancies, allogeneic blood and marrow transplantation or solid organ transplantation, who were refractory (86%) to, or intolerant (14%) of conventional antifungal therapy [132].Complete and partial responses to caspofungin were observed in 4 (5%) and 33 (40%) of these 83 patients.Treatment was well tolerated with minimal drug-related toxicity (clinical: 12%, laboratory: 14%).A non-comparative study on caspofungin for first-line therapy of invasive aspergillosis in 32 immunocompromised patients with haematological malignancies reported a complete or partial response in 56% of cases [133].IA-related mortality was 22%.Comparative clinical studies of first-line caspofungin therapy for patients with IA are needed.The Infectious Diseases Group of the European Organization for Research and Treatment of Cancer (EORTC-IDG) is currently conducting such a study.
Other echinocandins.In a non-comparative trial conducted in Japan in 19 patients with various forms of IA, success rates of micafungin were 60% to 67% [76].

Combination therapy
Given the poor prognosis of IA in immunocompromised patients, using combinations of antifungal drugs would seem a logical approach to improve patient's outcome.However, there are few clinical data supporting this approach.A few reports (consisting mainly of case reports and subgroups of patients in case series) of combination therapy in patients with IA suggest that liposomal amphotericin B plus caspofungin or voriconazole plus caspofungin may improve the outcome of patients with refractory IA [82,[134][135][136].A recent case-control study reported improved survival of patients with refractory, probable or proven IA treated with a combination of voriconazole and caspofungin compared with voriconazole monotherapy [137].In a multicentre, noncomparative salvage study in 53 adult patients with proven or probable invasive aspergillosis refractory or intolerant to first-line antifungal therapy the combination of caspofungin with amphotericin B, itraconazole, or voriconazole was successful in 50%, 43%, and 60% of cases, respectively [138].Success rates in patients with neutropenia or allogeneic haematopoietic stem cells transplant were 57% and 54%, respectively.The combinations were well tolerated, serious drug-related adverse events occurred in only 3% of cases.A case-control study compared the combination of voriconazole and caspofungin (n = 34) with a monotherapy with a lipid form of amphotericin B (n = 38) in solid organ transplant recipients with invasive aspergillosis [139].A significant reduction of 90-day mortality in the combination group was reported (26% vs 50%).In high-risk patients (ie dialysis, retransplantation, or disseminated aspergillosis) mortality was 23% in the combination group and 79% in the control group.However, large, prospective, randomised, comparative studies are needed to determine whether combination therapy is indeed superior to monotherapy.

Primary (ie upfront) therapy for invasive aspergillosis.
An intensive diagnostic work-up of the patient is recommended to determine the aetiology of infection in patients suspected of invasive aspergillosis.A recent report showing the limited efficacy of salvage antifungal therapy in patients suffering of IA has highlighted the importance of an appropriate upfront therapy [111].Voriconazole is first choice for upfront therapy of IA.Keeping in mind the interindividual variations of voriconazole pharmacokinetics, monitoring blood levels may be considered in patients who do not respond to therapy or in those suspected of developing drug-related adverse events.Alternative therapies are liposomal amphotericin B and amphotericin B deoxycholate.It is recommended to continue therapy until complete clinical and radiological resolution of infection.
Switch to oral itraconazole or voriconazole should be considered as soon as the patient shows signs of improvement and is able to take oral drugs as it will significantly reduce treatment costs.Monitoring of itraconazole blood levels is routinely recommended.

Rationale
Cancer patients at high risk of fungal infections are those with acute leukaemia or with allogenic hematopoietic stem cell transplantation (HSCT) who have prolonged and profound neutropenia or immunosuppression for graft-versushost disease.In contrast, patients with other haematological malignancies (myeloma, lymphomas, chronic myelogenous or lymphocytic leukaemia), solid tumours and recipients of an autologous HSCT are at low risk of invasive mycoses as neutropenia is typically of short duration and immunosuppression is rather moderate.Administration of empirical antifungal therapy has become a standard of care in patients with neutropenia and persistent fever despite treatment with broadspectrum antibiotics.The rationale for early administration of antifungal agents in these patients has been that clinically occult fungal infections (primarily due to Candida or Aspergillus) are a fre-quent autopsy finding in neutropenic patients and that persistent fever is often the only early sign of an invasive mycosis.

Clinical trials
The concept of empirical antifungal therapy is supported by the results of two pioneer, albeit not placebo-controlled, randomised studies conducted in the 1980s, which showed that the empirical use of amphotericin B in neutropenic patients with prolonged fever of undetermined origin reduced the incidence of fungal infections and fungal-related mortality [140,141].The benefit was primarily observed in patients who were severely neutropenic, who had not received antifungal prophylaxis, or who had a clinically documented infection [141].Based on the results of these initial studies, empirical administration of amphotericin B deoxycholate became standard therapy in these patients.

Empirical antifungal therapy in patients with neutropenia and persistent fever
During the last two decades, more than 20 randomised controlled clinical trials of empirical antifungal therapy have been performed in patients with neutropenia and persistent fever [8,9,46,117,[140][141][142][143][144][145][146][147][148][149][150][151][152][153][154][155][156][157][158].The vast majority of these studies have compared the efficacy and safety of various doses or formulations (ie conventional versus lipid formulations) of amphotericin B with that of an azole (ketoconazole or fluconazole) in an usually limited number of patients and therefore lacked power to detect small differences of efficacy or toxicity.We will therefore review the results of the most recent clinical trials that included a larger patient population and that used a composite score of several clinical and laboratory criteria to evaluate the response to therapy.In a non-inferiority (8 10%) study of 702 patients, liposomal amphotericin B was shown to be as effective as amphotericin B deoxycholate (50% vs 49%), but was associated with fewer breakthrough fungal infections (3.2% vs 7.8%), less infusion-related fever (17% vs 44%), chills and rigors (18% vs 54%) and less nephrotoxicity (19% vs 34%) [8].One should note, however, that the daily dose of amphotericin B deoxycholate (0.6 mg/kg/d) was moderate and that there was no mentioning of fluid loading to reduce amphotericin B deoxycholate toxicity.
Three clinical trials have compared the efficacy and safety of an azole (fluconazole, itraconazole or voriconazole) to that of amphotericin B deoxycholate or liposomal amphotericin B. In a study of 317 patients, favourable response rates occurred in 68% of the patients treated with intravenous fluconazole and in 67% of the patients treated with intravenous amphotericin B deoxycholate [153].Progressive or new fungal infections occurred in a similar proportion of patients in the two treatment groups (8% vs 6%, respectively), but adverse events occurred more often in patients treated with amphotericin B than in the fluconazole group (81% vs 13%, P = 0.001).Overall mortality and fungal infection-related mortality were similar in both treatment groups.In a large, open, non-inferiority (8 15%) study that included 384 patients, itraconazole was found to exhibit at least equivalent efficacy in an intention-to-treat analysis (47% vs 38%, difference: 9%, 95%CI: -0.8 to 19.5%) and significantly fewer drugrelated adverse events, including nephrotoxicity (5% vs 24%, P <0.001) than amphotericin B deoxycholate [155].Interestingly, the response rate to itraconazole was superior to that of amphotericin B deoxycholate in patients who had previously received antifungal prophylaxis (48% vs 35%, difference: 13%, 95%CI: 1.6 to 24.8%).Breakthrough fungal infections and mortality were similar in both treatment groups.In the third study, a non-inferiority (8 10%) multicentre trial that included 837 patients, the overall success rates were 26% for patients treated with voriconazole and 31% for patients treated with liposomal amphotericin B [156].However, voriconazole did not fulfill the statistical criteria for non-inferiority.Voriconazole was superior to liposomal amphotericin B for the prevention of breakthrough fungal infections (1.9% vs 5.0%, P = 0.02), especially in high-risk patients such as allogeneic transplants and patients with relapsed leukaemia.Fewer cases of infusion-related toxicity reactions (P <0.01) and of nephrotoxicity (P <0.001), but more cases of transient abnormal vision (P <0.001) and of hallucinations (P <0.001) occurred in the voriconazole group than in the liposomal amphotericin B group.
Finally, the results of a recent large, doubleblind, multicentre study that included 1123 patients showed that caspofungin was as effective as (overall response rates: 33.9% vs 33.7%), and better tolerated than lipososomal amphotericin B [158].Less infusion-related toxicity (35.1% vs 51.6%), nephrotoxicity (2.6% vs 11.5%), and premature discontinuation of study medication (10.3% vs 14.5%) occurred in patients treated with caspofungin than in those treated with liposomal amphotericin B. In addition, caspofungin was superior to liposomal amphotericin B for the successful outcome of baseline fungal infections (51.9% vs 25.9%) and for survival rates 7 days after therapy (92.6% vs 89.2%).
Clinical trials of empirical antifungal therapy for patients with neutropenia and persistent fever performed during the last two decades have not revealed a clear-cut superior antifungal agent in terms of efficacy.A consistent finding of many studies has been the increased risk of infusionrelated reactions and nephrotoxicity in patients treated with amphotericin B deoxycholate and to a lower extent with lipid formulations when compared with azoles (fluconazole, itraconazole or voriconazole) or echinocandins (caspofungin).However, given that it is as active as and substan-tially less expensive than most other antifungal drugs on the market, amphotericin B deoxycholate is recommended as first choice therapy for patients with neutropenia and persistent fever.Close monitoring of the appearance of adverse events in patients treated with amphotericin B deoxycholate is mandatory.Should contra-indications be present at baseline or develop during therapy (see footnote table 4), liposomal amphotericin B, caspofungin, itraconazole, or voriconazole are alternative treatment options.However, azoles should not be used for empirical antifungal therapy in patients receiv-  2 Few clinical data available on the use of azoles and echinocandins in neutropenic patients with documented systemic candidiasis.Few clinical data available on the use of caspofungin in neutropenic patients with documented invasive aspergillosis.Azoles and echinocandins are not active against zygomycetes.Due to the increased risk of azole-resistant Candida infections, azoles should not be used in patients receiving fluconazole prophylaxis.
Duration of empirical antifungal treatment for persistent fever during neutropenia : until recovery from neutropenia + resolution of clinical/microbiological/ radiological symptoms/signs suggestive of invasive fungal infection.Grade: CIII

GENERAL COMMENTS
Contraindications for the use of amphotericin B deoxycholate (Fungizone ® ): -Serum creatinine >150 micromol/L, Grade: CIII -History of toxicity which required discontinuation of therapy, Grade: CIII -Concomitant medication with nephrotoxic drugs (eg aminoglycosides, cyclosporin), Grade: CIII Toxicity criteria for discontinuation of therapy in patients receiving amphotericin B deoxycholate (Fungizone ® ): -Increase of serum creatinine: >2x baseline value (if normal at baseline), or >200 micromol/L (if baseline between upper normal limit and 150 micromol/L), Grade: CIII -Debilitating infusion-related fever and chills persisting for >2 days despite infusion over 24 hours and symptomatic treatment including paracetamol + opioid + anti-histaminic, Grade: CIII Monitoring of voriconazole blood levels should be considered if underdosing is suspected or in presence of clinical symptoms/signs suggestive of drug-related toxicity.Grade: CIII Monitoring of itraconazole blood levels routinely recommended.Grade: CIII Table 4 Empirical antifungal therapy in patients with neutropenia and persistent fever.

Swiss Medical Weekly
The many reasons why you should choose SMW to publish your research Med Wochenschr (1871-2000) Swiss Med Wkly (continues Schweiz Med Wochenschr from 2001) Editores Medicorum Helveticorum

.1. Absence of neutropenia, severe sepsis or septic shock or recent exposure to azoles. Flucona
zole is the drug of choice in haemodynamicallystable, non-neutropenic patients at low risk for azole-resistant Candida species.Alternative therapies are amphotericin B deoxycholate, caspofungin and voriconazole.1.2Recentexposure to azoles.A recent exposure to azoles increases the risk of an infection due to non-albicans Candida, such as C. glabrata or C. krusei, that exhibit reduced susceptibility (C.glabrata) or are resistant (C.krusei) to azoles.In patients in whom infections may be caused by azole-resistant non-albicans Candida species following previous therapy with azoles, amphotericin B deoxycholate or caspofungin are therefore rec- ® )* Liposomal Amphotericin B (AmBisome ® ) According to species identification to azoles 1 mg/kg/d i.v., Grade: BI 3 mg/kg/d i.v., Grade: BII and susceptibility testing or Caspofungin (Cancidas ® ) 70 mg i.v.Grade: CIII (1 st dose), then 50 mg/d i.v., Grade: BINeutropenic patientAmphotericin B deoxycholate (Fungizone ® )* Caspofungin (Cancidas ® ) 1 (see Table 2B) 70 mg i.v.According to species identification, 1 mg/kg/d i.v., Grade: CIII (1 st dose, then 50 mg/d i.v., Grade: CIII * See general comments in the footnote of table 4.

Species Susceptibility testing First choice Alternatives Switch to oral medication
S: susceptible; S-DD: susceptible-dose dependent; R: resistant.The evidence that in vitro susceptibility of C. glabrata to fluconazole predicts clinical response is lacking.* See general comments in the footnote of table 4.
infections caused by C. albicans, C. tropicalis or C. parapsilosis, fluconazole is the drug of choice.Switch to oral therapy is recommended