How to improve the design of trials of antifungal prophylaxis among neutropenic adults with acute leukemia.

The risk for invasive fungal infections in patients with acute leukemia is generally low (4%-8%), and the routine use of fungal prophylaxis is not warranted except in specific high-risk groups that should be identified among this population. In a prophylactic study with a new agent, fluconazole or itraconazole oral solution represent good choices for the comparator because they are proven better than placebo or oral nonabsorbable antifungal agents in reducing the risk of invasive fungal infections in patients with acute leukemia. Because prophylaxis is most valuable when the risk of infection is high, patients with well-understood risk factors (severe mucosal disruption caused by chemotherapy, impaired cell-mediated immunity caused by steroids or fludarabine, use of a central venous catheter, and colonization by Candida species) should be selected. The end points for antifungal prophylactic trials should focus on proven and probable invasive fungal infections. Superficial and mucosal fungal infections do not represent a primary end point for these studies. Poor compliance should be considered as an interruption of treatment due to side effects and should be included in the criteria for failure. Fungus-related mortality should be evaluated as a failure of prophylaxis, whereas overall mortality may be influenced by many other cofactors. Differences in gastrointestinal toxicity of antifungal agents may limit the use of double-blind designs in some situations.