Abstract
The original trials of empiric intravenous amphotericin-B in the 1980s failed to prove conclusively its efficacy in the treatment of febrile neutropenia. Despite that, all subsequent studies of the therapy of presumed, possible, probable and proven invasive aspergillosis have assumed that this drug, either as deoxycholate or in lipid-based form, is the gold standard treatment against which all newcomers should be compared. This has led to a series of further inconclusive randomized controlled trials of empiric therapy as a result of which the most we can say is that nearly all new drugs are less toxic but also no more effective than amphotericin-B deoxycholate. The toxicity of the non-lipid formulation of this drug should have led us to withdraw it from both RCTs and routine clinical practice some years ago in view of the increasing evidence of equivalent efficacy and lower toxicity of other agents including lipid amphotericin formulations.
Recent studies of the use of newer diagnostic techniques (i.e., CT and serology) reinforce the need to abandon the empiric trial approach in which we have repeatedly shown lack of superiority in the treatment of an infection which most patients do not have. Even in the small number of trials of the therapy of proven or probable invasive aspergillosis, results have been inconclusive or at best confusing in trying to find a better option than amphotericin-B. The trial of voriconazole versus amphotericin-B deoxycholate for this indication is a model for study for all those interested in the difficulties of designing trials which lead to convincing results.
Effective prophylaxis trials and their analyses began by following a more rational pathway, first showing convincingly that fluconazole reduced the risk of C. albicans systemic infection in transplant patients. Unfortunately the widespread faith in the ability of this drug to prevent a wider range of systemic fungal infections in a wider range of patients is simply not supported by the data from many subsequent single trials and meta-analyses. This attachment to fluconazole has been mirrored by unwillingness to accept the evidence that itraconazole is superior in prophylaxis to fluconazole which is inactive against Aspergillus spp. In this case the trials have not told us enough because we have not believed the results. Results of trials of extended range azoles such as posaconazole are interesting but there are insufficient data to claim that posaconazole is superior to itraconazole.
The progress in therapy and prophylaxis of systemic fungal infection has been unsatisfactory and slow. A new approach is needed for the design of clinical trials for these indications. There is good evidence that supportive investigations should now be used routinely in clinical practice and trials to increase certainty about the presence of invasive infection, to limit unnecessary use of expensive and toxic drugs and to improve analysis of efficacy of old and new antifungal agents.