Antifungal prophylaxis for severely neutropenic chemotherapy recipients: a meta analysis of randomized-controlled clinical trials

Long-term antifungal prophylaxis in high-risk hematopoietic stem cell transplant recipients

The risks for invasive fungal infections, particularly mould infections such as invasive aspergillosis, among hematopoietic stem cell transplant (HSCT) recipients are linked to the duration and severity of myelosuppression and immunosuppression. Strategies to prevent invasive fungal infections have focused primarily on the use of orally administered azole antifungal agents during the neutropenic period rather than on the more prolonged post-engraftment period. The major limitations of these studies included the heterogeneity among the subjects studied for fungal infection risk factors, the agents administered, the dosing, and duration of prophylaxis. More recent studies have attempted to examine the efficacy of antifungal prophylaxis strategies among allogeneic HSCT recipients to day 100 and beyond. It is clear that a variety of products have efficacy in preventing invasive candidiasis, including imidazole and triazole antifungals, low-dose amphotericin B, and the echinocandin, micafungin; however, only the extended spectrum azole, itraconazole, has been shown to impact the incidence of proven invasive aspergillosis. Other extended spectrum azole antifungal agents, voriconazole and posaconazole, are being studied as long-term prophylaxis in high-risk HSCT recipients. While clinical trials have suggested that a duration of prophylaxis against moulds of six months or more may be required, it remains unclear if this is required in all cases. The prophylactic efficacy over time may be linked to the degree of immunosuppression as measured by markers such as the numbers of circulating CD4 T lymphocytes. Concerns about selection for resistant moulds among long-term recipients of these drugs are emerging. The cumulative experience to date suggests that long-term antifungal chemoprophylaxis is feasible and effective when applied in defined circumstances. The concerns about treatment-related toxicities, resistance, and costs are valid.

Recent developments in the management of invasive fungal infections in patients with hematological malignancies

Despite recent advances in the last decade, invasive fungal infections are still associated with a high morbidity and mortality. Invasive fungal infections constitute severe infectious complications in patients with hematological malignancies receiving myelosuppressive chemotherapy or sustained immunosuppression after allogeneic transplant regimens. Following a long period of stagnation, considerable progress has been made during the last 5 years in non-culture-based diagnostics and in the treatment of invasive fungal infections. This review highlights recent developments in the epidemiology, diagnosis, and treatment in the context of state-of-the-art management of invasive fungal infections in cancer patients.

Intrapulmonary pharmacokinetics and pharmacodynamics of itraconazole and 14-hydroxyitraconazole at steady state

We determined the steady-state intrapulmonary pharmacokinetic and pharmacodynamic parameters of orally administered itraconazole (ITRA), 200 mg every 12 h (twice a day [b.i.d.]), on an empty stomach, for a total of 10 doses, in 26 healthy volunteers. Five subgroups each underwent standardized bronchoscopy and bronchoalveolar lavage (BAL) at 4, 8, 12, 16, and 24 h after administration of the last dose. ITRA and its main metabolite, 14-hydroxyitraconazole (OH-IT), were measured in plasma, BAL fluid, and alveolar cells (AC) using high-pressure liquid chromatography. Half-life and area under the concentration-time curves (AUC) in plasma, epithelial lining fluid (ELF), and AC were derived using noncompartmental analysis. ITRA and OH-IT maximum concentrations of drug (C(max)) (mean +/- standard deviation) in plasma, ELF, and AC were 2.1 +/- 0.8 and 3.3 +/- 1.0, 0.5 +/- 0.7 and 1.0 +/- 0.9, and 5.5 +/- 2.9 and 6.6 +/- 3.1 microg/ml, respectively. The ITRA and OH-IT AUC for plasma, ELF, and AC were 34.4 and 60.2, 7.4 and 18.9, and 101 and 134 microg. hr/ml. The ratio of the C(max) and the MIC at which 90% of the isolates were inhibited (MIC(90)), the AUC/MIC(90) ratio, and the percent dosing interval above MIC(90) for ITRA and OH-IT concentrations in AC were 1.1 and 3.2, 51 and 67, and 100 and 100%, respectively. Plasma, ELF, and AC concentrations of ITRA and OH-IT declined monoexponentially with half-lives of 23.1 and 37.2, 33.2 and 48.3, and 15.7 and 45.6 h, respectively. An oral dosing regimen of ITRA at 200 mg b.i.d. results in concentrations of ITRA and OH-ITRA in AC that are significantly greater than those in plasma or ELF and intrapulmonary pharmacodynamics that are favorable for the treatment of fungal respiratory infection.

Antimicrobial prophylaxis in hematopoietic stem cell transplant recipients: heterogeneity of current clinical practice

Antimicrobial agents are commonly used after hematopoietic stem cell transplant (HSCT) to prevent bacterial, viral and fungal infections. A pharmacy practice survey was undertaken to evaluate prevailing practices. The 31 centers evaluated transplanted over 3400 patients in 2001. Over half used bacterial prophylaxis; all with fluoroquinolones. A significantly higher proportion (90-100%) used fungal and viral prophylaxis. Most centers used fluconazole for fungal prophylaxis, but the dose used varied from 400 mg (the recommended dose) to 100 mg. Itraconazole and amphotericin preparations were used by some centers for allograft recipients because of their activity against aspergillosis. Most centers used brief viral prophylaxis for autograft recipients aimed at preventing HSV reactivation. Viral prophylaxis for allograft recipients was usually much more prolonged, reflecting concern over cytomegalovirus infections. Overall, there was significant deviation from recommended guidelines in many of the practices. Our survey suggests that substantial variation exists among transplant centers in their approach to antimicrobial prophylaxis after HSCT. This probably stems from the lack of definitive studies and strong recommendations in several areas, availability of newer agents that have not been adequately studied in the HSCT setting, and a desire to improve outcome before definitive studies are available for newer agents, a process that could take several years.

Optimum management of pediatric patients with fever and neutropenia

Fever with neutropenia is a common clinical problem in patients receiving cancer treatment. Prevention and optimum management of infectious complications is critical to the overall success of cancer therapy. This article provides an overview of the current status of this evolving subject. While the basic principles of rapid institution of broad spectrum antibiotics, early intervention with empiric antifungal therapy and continuation of antimicrobials during period of risk are unlikely to change, there is increasing interest in titrating this aggressive approach based on the projected risk of the development of a serious invasive infection. Oral antibiotic therapy and outpatient management are currently being studied in pediatric oncology patients, but even when successful these alternatives to the traditional "in hospital, parenteral antibiotic therapy" approach are unlikely to be applicable in all patient populations and clinical settings. While there is no replacement for clinical acumen and careful monitoring, judicious use of diagnostic resources such as blood cultures and imaging studies is a key component of optimum care. Selection of empiric antibiotics based on ongoing monitoring of antimicrobial susceptibility patterns is emphasized.

Choices aplenty: antifungal prophylaxis in hematopoietic stem cell transplant recipients

The incidence of invasive fungal infection (IFIs) in hematopoietic stem cell transplantation (HSCT) recipients ranges from 10 to 25% with an overall case fatality rate of up to 70-90%. Candida and Aspergillus genera remain the two most common pathogens. Although fluconazole prophylaxis in this population has been moderately effective in reducing mortality due to invasive candidiasis, this agent does not have activity against invasive aspergillosis (IA) and other mould. Several new agents such as voriconazole and caspofungin have enhanced potency and broad-spectrum antifungal activity and show promising results against yeasts and filamentous fungi when given as therapy and as chemoprophylaxis. Further, new diagnostic tools to detect circulating fungal antigens in biological fluids and PCR-based methods to detect species or genus-specific DNA or RNA have been developed. Incorporating these techniques along with clinical criteria appear to improve the accuracy of preclinical diagnosis of IFIs. Such approaches may alter the current treatment strategy from prophylaxis to pre-emptive therapy, thereby potentially decreasing cost and toxicity in high-risk patients.

Attenuation of itraconazole fungicidal activity following preexposure of Aspergillus fumigatus to fluconazole

Fluconazole (FLC), a triazole with limited activity against Aspergillus species, is frequently used as prophylaxis in leukemia patients and bone marrow transplant recipients. Prior FLC use has been associated with an increasing incidence of invasive aspergillosis in these patients. We hypothesized that prior exposure of Aspergillus fumigatus to FLC could result in altered in vitro susceptibility of this fungus to other, more active triazoles. Thus, we performed serial passages of conidia of 10 clinical isolates of A. fumigatus (all itraconazole [ITC] susceptible) on FLC-containing yeast agar glucose plates. The MICs and minimal fungicidal concentrations (MFCs) of amphotericin B, FLC, ITC, and voriconazole (VRC) for A. fumigatus conidia were measured following four passages on FLC-containing medium according to the National Committee for Clinical Laboratory Standards microdilution method. Serial passages on FLC-containing plates resulted in a fourfold increase in the MFCs (but not the MICs) of ITC for nine isolates. The attenuated ITC fungicidal activity against A. fumigatus following FLC preexposure was medium independent and was also observed against FLC-preexposed A. fumigatus hyphae with the viability staining FUN-1 dye. Moreover, FLC preexposure of A. fumigatus conidia resulted in an analogous increase in the MFCs (but not the MICs) of VRC. Our findings suggest that preexposure of A. fumigatus to FLC attenuates the in vitro fungicidal activity of subsequent ITC use against it. This phenotypic adaptation is not captured by a routine MIC determination but requires MFC measurement. The in vivo significance of this in vitro phenomenon requires further investigation.

Novel approaches to antifungal prophylaxis

Antifungal prophylaxis represents a significant advance in the management of patients at risk from fungal infections in a variety of settings. Identification of patients at the highest risk and the utilisation of safe and effective drugs maximises the benefits of prophylaxis. Situations in which antifungal prophylaxis has been shown to be useful are bone marrow transplantation, liver and lung transplantation, surgical and neonatal intensive care units, secondary prophylaxis of fungal infections associated with HIV and neutropenia associated haematological malignancies and their treatment. New antifungal agents, such as the echinocandins and the new azoles, are available and have a potential role in antifungal prophylaxis. Future studies should evaluate which strategy is more useful; prophylaxis or pre-emptive therapy.

Levels of total fungus and Aspergillus on a pediatric hematopoietic stem cell transplant unit

The purpose of this descriptive study was to determine the levels of total fungus (TF) and Aspergillus in a pediatric hematopoietic stem cell transplant (HSCT) unit. One hundred twenty air samples and 120 floor samples were collected from the same locations in 10 patient rooms and bathrooms for 4 consecutive days. The count in colony-forming units of TF and Aspergillus from each of the samples was measured by the institution's mycology laboratory. Means, standard deviations, minimum values, and maximum values were determined for levels of TF and Aspergillus from different locations and on different days in the air and on the floor. Determination of a mean value of TF and Aspergillus for each room allowed for analysis of mean values of TF and Aspergillus for sample category, room side, room type, and room status. After visual examination of the mean values for the air samples collected, it was determined that the TF and Aspergillus in the air were less than the institution's acceptable air baseline standard. t tests and analysis of variance were used to verify the findings.

Invasive aspergillosis in patients with hematologic malignancies

Invasive aspergillosis is an increasingly common and often fatal opportunistic fungal infection in patients with hematologic malignancies. Prolonged and profound neutropenia remains a key risk factor for the development of invasive aspergillosis. However, qualitative deficiencies in host immune responses resulting from prolonged corticosteroid therapy, graft-versus-host disease, and cytomegalovirus infection are important risk factors for the recurrence and progression of Aspergillus infections after bone marrow recovery. Early diagnosis of invasive aspergillosis remains a challenge, and few tools are available for monitoring its course once the diagnosis is established. Even with the recent introduction of new antifungal therapies, mortality in patients with invasive aspergillosis remains high, and uniformly effective prophylaxis or preemptive therapeutic strategies are lacking. Strategies such as combination antifungal therapy and immunotherapy often are used as first-line treatment approaches in patients with documented invasive aspergillosis despite a paucity of clinical trial data. Recent advances in our understanding of the epidemiology, pathogenesis, and treatment of invasive aspergillosis in patients with hematologic malignancies are reviewed. The problems and controversies associated with defining optimal treatment strategies for invasive aspergillosis in this heavily immunocompromised population are highlighted.

The role of itraconazole in preventing and treating systemic fungal infections in immunocompromised patients

The increasing use of immunosuppressive chemotherapy and allogeneic transplants for haematological malignancies has increased the number of patients at risk of systemic fungal infections (SFIs). A number of antifungal agents are now available. This paper reports the deliberations of an expert panel that considered the role of itraconazole. It concluded that itraconazole has an important role in the prophylaxis and treatment of SFIs whether proven or not.

Amphotericin B lipid complex as prophylaxis of invasive fungal infections in patients with acute myelogenous leukemia and myelodysplastic syndrome undergoing induction chemotherapy

BACKGROUND

The optimal antifungal prophylactic regimen for patients with acute myelogenous leukemia (AML) or high-risk myelodysplastic syndrome (MDS) undergoing induction chemotherapy has yet to be identified. A prospective historical control study evaluated the efficacy and safety of amphotericin B lipid complex (ABLC) in this patient population.

METHODS

Newly diagnosed patients with AML or high-risk MDS who were undergoing induction chemotherapy received prophylactic ABLC 2.5 mg/kg intravenously 3 times weekly. This treatment group was compared with a historical control group that had similar baseline characteristics and received prophylactic liposomal amphotericin B (L-AmB) 3 mg/kg 3 times weekly. The primary endpoint was the incidence of documented or suspected fungal infections during and up to 4 weeks after cessation of prophylaxis. Reported adverse events were used to assess tolerability.

RESULTS

The overall efficacy of antifungal prophylaxis was similar in patients who received ABLC and patients who received L-AmB (P=0.95). Among 131 ABLC-treated patients and 70 L-AmB-treated patients who were assessed for efficacy and safety, 49% of patients in each group completed therapy without developing a documented or suspected fungal infection. Documented fungal infections occurred in 5% of ABLC-treated patients and in 4% of L-AmB-treated patients. Alternative antifungal strategies were required because of persistent fever or pneumonia of unknown pathogen in 28% and 32% of ABLC-treated and L-AmB-treated patients, respectively. Grade 3 and 4 adverse events, therapy discontinuations due to adverse events, and survival rates also were similar between treatment groups.

CONCLUSIONS

ABLC and L-AmB appeared to have similar efficacy and were tolerated well as antifungal prophylaxis in patients with AML and high-risk MDS who were undergoing induction chemotherapy.

Itraconazole prevents invasive fungal infections in neutropenic patients treated for hematologic malignancies: evidence from a meta-analysis of 3,597 patients

PURPOSE

Efficacy of antifungal prophylaxis has not yet been convincingly proven in numerous trials of various antifungals. New evidence and the anti-Aspergillus efficacy of itraconazole prompted a new look at the data for the prevention of invasive fungal infections.

PATIENTS AND METHODS

Randomized, controlled studies with itraconazole for antifungal prophylaxis in neutropenic patients with hematologic malignancies were identified from electronic databases and hand searching.

RESULTS

Thirteen randomized trials included 3,597 patients who were assessable for invasive fungal infections. Itraconazole reduced the incidence of invasive fungal infection (mean relative risk reduction, 40% +/- 13%; P =.002), the incidence of invasive yeast infections (mean, 53% +/- 19%; P =.004) and the mortality from invasive fungal infections (mean, 35% +/- 17%; P =.04) significantly. The incidence of invasive Aspergillus infections was only reduced in trials using the itraconazole cyclodextrine solution (mean, 48% +/- 21%; P =.02) and not itraconazole capsules (mean, 75% +/- 73% increase; P =.3). The overall mortality was not changed. Adverse effects were rare, hypokalemia was noted in three studies, and a higher rate of drug discontinuation was found in trials that compared itraconazole cyclodextrine solution to a control without cyclodextrine. The effect of prophylaxis was clearly associated with a higher bioavailable dose of itraconazole.

CONCLUSION

Antifungal prophylaxis with itraconazole effectively prevents proven invasive fungal infections and-shown for the first time for antifungal prophylaxis-reduces mortality from these infections and the rate of invasive Aspergillus infections in neutropenic patients with hematologic malignancies. Adequate doses of the oral cyclodextrine solution (at least 400 mg/d) or i.v. formulations (200 mg/d) of itraconazole are necessary for these effects.

Invasive cutaneous fungal infections requiring radical resection in cancer patients undergoing chemotherapy

Invasive fungal infections have emerged as a significant problem in patients with cancer with the development of better systemic therapies for malignancy and more effective antibacterial agents. The currently available world published medical literature was reviewed on invasive fungal infections in cancer patients with specific attention devoted to the multidisciplinary role of surgery in refractory cutaneous cases. Infections can develop on the forearm where peripheral intravenous catheters had been inserted in cancer patients undergoing cytotoxic chemotherapy. Curative intent begins with systemic contemporary anti-fungal therapy. Following resolution of neutropenia, patients may require radical surgical debridement with negative margins of resection for complete eradication of the fungal infection. Although invasive fungal infections refractory to antifungal systemic therapy in immunocompromised patients undergoing chemotherapy are a rare event, it is critical for surgeons and other multidisciplinary clinicians to recognize these potentially life-threatening infections that may necessitate radical surgical resection for cure.

Antifungal prophylaxis with azole derivatives

In recent years, several reports have underlined the increasing role of fungal infections as a cause of morbidity and mortality in hospitalised patients. For this reason, and also in light of the high mortality rate associated with these infections, chemoprophylaxis has been advocated by several authors. The available evidence suggests that both fluconazole and itraconazole are able to decrease candida colonisation and infection, when compared with placebo or with nonabsorbable antifungals. Data seem also to suggest that a decrease in fungus-related mortality can be achieved with prophylaxis, although with little effect on overall mortality, probably because of the importance of severe underlying diseases. Itraconazole proved to be effective in the prevention of fungal infections, including invasive aspergillosis, although with increased incidence of side-effects, often leading to treatment discontinuation. The other side of the coin is that antifungal prophylaxis might have untoward effects, such as the selection of triazole-resistant Candida strains or the induction of resistance. In addition, some authors have suggested that the use of triazoles might modulate the pattern of infecting organisms in cancer patients, increasing the risk of both aspergillosis and bacteremia. In conclusion, antifungal prophylaxis with triazole antifungals should be used with caution, only in patients at high risk for invasive fungal infections. These include allogeneic bone marrow transplant patients (especially those with mismatched or unrelated donors), acute myeloid leukaemia patients treated with high-dose cytarabine (C-ara), very-low-birth-weight infants, patients with chronic granulomatous disease, and high-risk surgical and intensive-care unit patients.

Management of candidiasis Management of Candida species infections in critically ill patients

Invasive candidiasis is a feared infection with mortality similar to that of septic shock (40-60%). Improved knowledge of its pathophysiology and the availability of new compounds for antifungal therapy and prophylaxis have contributed to improving the prognosis of severe candidal infections among immunosuppressed patients at the possible cost of the emergence of non-albicans strains of candida with lower susceptibility to azoles. This review focuses on the management of invasive deep-seated candidiasis in critically ill, non-immunocompromised patients. We discuss antifungal use, indications, potential benefit, and main secondary effects. Prevention strategies include pre-emptive antifungal therapy and azole-based prophylaxis. For patients at lower initial risk, pre-emptive therapy should be based on a management strategy that takes into account the presence of definite risk factors and the dynamics of candida colonisation. Among critically ill patients, azole prophylaxis is effective and is not associated with acquisition of resistance; it must be restricted to highly selected groups of patients at high risk only.

Epidemiology of Candida species infections in critically ill non-immunosuppressed patients

A substantial proportion of patients become colonised with Candida spp during hospital stay, but only few subsequently develop severe infection. Clinical signs of severe infection manifest early but lack specificity until late in the course of the disease, thus representing a particular challenge for diagnosis. Mostly nosocomial, invasive candidiasis occurs in only 1-8% of patients admitted to hospitals, but in around 10% of patients housed in intensive care units where it can represent up to 15% of all nosocomial infections. We review the epidemiology of invasive candidiasis in non-immunocompromised, critically ill patients with special emphasis on disease trends over time, pathophysiology, diagnostic approach, risk factors, and impact. Recent epidemiological data suggesting that the emergence of non-albicans candida strains with reduced susceptibility to azoles, previously linked to the use of new antifungals for empiric and prophylactic therapy in immunocompromised patients, may not have occurred in the critically ill. Management of invasive candidiasis in these patients will be addressed in the December issue of The Lancet Infectious Diseases.

The global epidemiology of invasive Candida infections—is the tide turning?

Invasive candidiasis is a condition of major medical importance. Its incidence has increased dramatically over the last 50 years, reflecting increasingly interventional standards of medical care. Candida spp. are regularly reported to be the fourth commonest cause of bloodstream infection, and it is perceived that the incidence of invasive Candida spp. infections continues to increase. The global disease burden of invasive Candida spp. infections is difficult to quantify because of wide geographic variation. Data originating from the United States indicate that mortality from candidiasis has been falling since 1989. Data from several locations have shown that the dramatic increases in Candida spp. bloodstream infections seen during the 1980s were not sustained through the 1990s. Some authors have reported a decreasing incidence. The contribution of non-albicans Candida spp. to invasive infection is rising. Invasive infections with Candida spp. continue to represent a major economic burden, increasing both mortality and morbidity in an already expensive group of hospital patients. There remains much scope for ongoing and future research into the epidemiology and basic disease processes underlying these infections.

Infections in patients with cancer undergoing chemotherapy: aetiology, prevention, and treatment

Patients with cancer who are undergoing chemotherapy are highly susceptible, especially if neutropenic, to almost any type of bacterial or fungal infection. These infections cause substantial morbidity and mortality. Prophylactic use of antibiotics should be avoided, however, since this practice is associated with a risk of emergence of resistant bacteria and it does not lower the risk of death. However, chemoprophylaxis has a role for candidal fungal infections. Because infection in a neutropenic host can be rapidly fatal if not treated, the empirical administration of broad-spectrum intravenous antibiotics is generally indicated for these patients, and the local frequencies, susceptibility, and resistance patterns of various pathogens must be taken into account. Once therapy has been initiated, changes in antibiotic regimens during the first 5 days are useless unless the patient's clinical condition deteriorates substantially. The treatment of invasive fungal infections is particularly difficult. Many unsolved questions remain, and studies are proposed here that may shed light on these issues.

Candidemia in critically ill patients: difference of outcome between medical and surgical patients

OBJECTIVE

Candidemia is increasingly encountered in critically ill patients with a high fatality rate. The available data in the critically ill suggest that patients with prior surgery are at a higher risk than others. However, little is known about candidemia in medical settings. The main goal of this study was to compare features of candidemia in critically ill medical and surgical patients.

DESIGN

Ten-year retrospective cohort study (1990-2000).

SETTING

Medical and surgical intensive care units (ICUs) of a teaching hospital.

PATIENTS

Fifty-one patients with at least one positive blood culture for Candida species.

MAIN RESULTS

Risk factors were retrieved in all of the patients: central venous catheter (92.1%), mechanical ventilation (72.5%), prior bacterial infection (70.6%), high fungal colonization index (45.6%). Candida albicans accounts for 55% of all candidemia. The overall mortality was 60.8% (85% and 45.2% in medical and surgical patients, respectively). Independent factors associated with survival were prior surgery (hazard ratio [HR] =0.25; 0.09-0.67 95% confidence interval [CI], p<0.05), antifungal treatment (HR =0.11; 0.04-0.30 95% CI, p<0.05) and absence of neutropenia (HR =0.10; 0.02-0.45 95% CI, p<0.05). Steroids, neutropenia and high density of fungal colonization were more frequently found among medical patients compared to surgical ones.

CONCLUSIONS

Candidemia occurrence is associated with a high mortality rate among critically ill patients. Differences in underlying conditions could account for the poorer outcome of the medical patients. Screening for fungal colonization could allow identification of such high-risk patients and, in turn, improve outcome.

Amphotericin B versus fluconazole for controlling fungal infections in neutropenic cancer patients

BACKGROUND

Systemic fungal infection is considered to be an important cause of morbidity and mortality in cancer patients, particularly those with neutropenia. Antifungal drugs are often given prophylactically, or to patients with persistent fever.

OBJECTIVES

To compare the effect of fluconazole and amphotericin B on morbidity and mortality in patients with cancer complicated by neutropenia.

SEARCH STRATEGY

MEDLINE and Cochrane Library (November 2001). Letters, abstracts, and unpublished trials. The industry and authors were contacted.

SELECTION CRITERIA

Randomised trials comparing fluconazole with amphotericin B.

DATA COLLECTION AND ANALYSIS

Data on mortality, invasive fungal infection, colonisation, use of additional (escape) antifungal therapy and adverse effects leading to discontinuation of therapy were extracted by both authors independently.

MAIN RESULTS

Sixteen trials (3760 patients, 341 deaths) were included. In 3 large 3-armed trials, results for amphotericin B were combined with results for nystatin in a "polyene" group. Because nystatin is an ineffective drug in these circumstances, this approach creates a bias in favour of fluconazole. Furthermore, most patients were randomised to oral amphotericin B, which is poorly absorbed and poorly documented. It was unclear whether there was overlap among the "polyene" trials. We were unable to obtain any information to clarify these issues from the trial authors or from Pfizer, the manufacturer of fluconazole. There were no significant differences in effect between fluconazole and amphotericin B, but the confidence intervals were wide. More patients dropped out of the study when they received amphotericin B, but as none of the trials were blinded, decisions on premature interruption of therapy could have been biased. Furthermore, amphotericin B was rarely given under optimal circumstances, with premedication to reduce infusion-related toxicity, slow infusion, and with potassium and magnesium supplements to prevent nephrotoxicity.

REVIEWER'S CONCLUSIONS

Amphotericin B had been disfavoured in several of the trials through their design or analysis. Since intravenous amphotericin B is the only antifungal agent for which there is good evidence suggesting an effect on mortality and is considerably cheaper than fluconazole, it should be preferred.