Australasian Society for Infectious Diseases

Guidelines for use of antifungal agents in treatment of invasive fungal infection (July 2003)

EMPIRICAL ANTIFUNGAL THERAPY IN FEBRILE NEUTROPENIC PATIENTS

Lipid Formulations

All the comparative studies of lipid formulations in fever and neutropenia have used differences in toxicity as the primary endpoint, with the exception of the study by Walsh et. al. [20], which compares the efficacy of amphotericin B versus liposomal amphotericin B as the primary endpoint.

Two direct comparisons of amphotericin B lipid complex and liposomal amphotericin B have been undertaken. Both focus on differences in toxicity rather than efficacy [8, 22]. Three studies [9, 20, 23] have compared amphotericin B to the lipid products, liposomal amphotericin B or amphotericin B lipid complex. These three studies all detected less nephrotoxicity with liposomal amphotericin B in comparison to amphotericin B [20, 23] or amphotericin B lipid complex [8]. Whether this difference in renal toxicity translates into significant clinical benefit remains open to debate. No survival benefit for a lipid formulation in the treatment of fever and neutropenia has been demonstrated. Whilst amphotericin B colloidal dispersion has less nephrotoxicity compared to amphotericin B, infusional reactions, especially the rate of hypoxic reactions were of concern [9]. In patients at highest risk of IFI, i.e. patients with prolonged neutropenia, liposomal amphotericin B at 3 mg/kg/day may be superior to amphotericin B or liposomal amphotericin B at 1 mg/kg/day [20, 23].

Triazoles

Triazoles studied in the setting of fever and neutropenia have been fluconazole, itraconazole and voriconazole. All have been compared with amphotericin B [14, 24, 25] or liposomal amphotericin B [26]. Two studies have compared fluconazole 400 mg daily to amphotericin B 0.5 mg/kg/day [24, 25]. Only one was powered to detect a difference in efficacy and showed a similar efficacy for both agents [24]. Adverse events were more common with amphotericin B but there was no difference in the rate of hepatic dysfunction. However, these studies were in patient populations probably at a lower risk for IFI than those described above [9, 20]. There were few allogeneic HSCT recipients and patients had not received prior fluconazole prophylaxis. An open-label, randomised comparison of itraconazole and amphotericin B in haematology patients, excluding allogeneic HSCT recipients, showed similar rates of death and breakthrough fungal infections [14]. Less toxicity, including nephrotoxicity was observed in the itraconazole arm. After 2 days of intravenous itraconazole the level was therapeutic (>250 ng/ml) in 97% of patients and these levels were maintained on oral itraconazole when substituted for intravenous as early as day seven.

The largest study of a triazole in patients with fever and neutropenia, an open label study in 837 patients, compared voriconazole and liposomal amphotericin B [26]. The overall success rates were 26% with voriconazole and 31% with liposomal amphotericin B. Voriconazole did not fulfil the protocol definition of non-inferiority which was an overall response rate within 10 percentage points of liposomal amphotericin B at the 95% confidence interval. There were fewer documented breakthrough fungal infections with voriconazole. This benefit was most marked in patients considered as high risk for fungal infection. No survival benefit was observed although survival was assessed at 7 days after completing therapy. At each of the 5 composite end-points other than breakthrough fungal infections the results favoured liposomal amphotericin B over voriconazole. However, more patients receiving voriconazole were withdrawn from the study prematurely due to concerns regarding lack of efficacy with ongoing fever. The open study design may result in biases. There were fewer infusional reactions and less nephrotoxicity with voriconazole. The incidence of hepatic toxicity was similar but voriconazole recipients had more transient visual disturbances and hallucinations. These side effects of voriconazole while acceptable when treating IFI may be less acceptable in patients receiving empiric therapy.

These studies demonstrate that triazoles may be as effective as amphotericin B or liposomal amphotericin B in empiric therapy of patients with fever and neutropenia. When applying these results, the selective nature of the inclusion criteria used should be considered. For example, although itraconazole performed as well as amphotericin B [14], it should be noted that allogeneic HSCT recipients were excluded from this study. Similarly the study of Winston et al. [24] had few allogeneic transplants and included no patients receiving fluconazole prophylaxis. The role of voriconazole in the treatment of patients with fever and neutropenia requires further delineation but it is an alternative to amphotericin B especially in patients at high risk of IFI.

Prolonged Amphotericin B Infusion

A 24 hour infusion of amphotericin B has been compared to a 4 hour infusion at a dose of approximately 1 mg/kg/day on starting antifungal therapy [27]. Continuous infusion had less overall infusional toxicity and nephrotoxicity and fewer deaths during treatment and at 3 months after end of treatment. Whilst this report is encouraging, this study was powered to detect changes in toxicity not efficacy. Also while the difference in death rates between the 2 arms was significant, there were no data on cause of death.

Furthermore, the effectiveness of the 24 hour infusion strategy in protecting against further nephrotoxicity with established nephrotoxicity is unknown. More research is required to evaluate this mode of administration.