We do not dispute the efficacy of fluconazole for prevention of Candida infections in transplant recipients (1, 2). However, the usefulness of fluconazole for antifungal prophylaxis in recipients of allogeneic hematopoietic stem-cell transplants has recently been compromised by the increasing incidence of Aspergillus and certain fluconazole-resistant Candida infections. Furthermore, while the Seattle study reported improved long-term survival in patients receiving prophylactic fluconazole after allogeneic hematopoietic stem-cell transplantation (3), these results were not seen in the study by Goodman and colleagues and in other trials of fluconazole prophylaxis (1, 2). It is also uncertain that the improved long-term survival in the Seattle study can be attributed to a reduction of fungal infections by fluconazole administered many years earlier rather than to many other factors and therapeutic interventions. In our study, the favorable trend of a reduction in death from fungal infection in patients taking prophylactic itraconazole probably did not achieve statistical significance because of the study sample size.

Some of the issues of trial design and study end points raised by Drs. Powers and Higgins and Dr. Mossad were discussed in our article. It is not appropriate to compare the low rate of fungal infection with fluconazole prophylaxis in the trial by Van Burik and colleagues (2.4%) (4) and the higher rate of fungal infection with fluconazole in our study (25%). Van Burik and colleagues compared fluconazole with micafungin for antifungal prophylaxis only during the period of neutropenia (18 days) immediately after transplantation in a group that included recipients of autologous as well as allogeneic hematopoietic stem-cell transplants. Recipients of autologous transplants have a considerably lower risk for fungal infection, while the risk for invasive fungal infections among recipients of allogeneic transplants is now much greater after engraftment (usually in association with increased immunosuppression for graft-versus-host disease) than during the early period of neutropenia after transplantation. Similarly, we believe that it is not appropriate to consider all deaths as failure of antifungal prophylaxis. Relapse of malignant disease, graft-versus-host disease, hemorrhage, and other noninfectious factors frequently cause death after hematopoietic stem-cell transplantation and would probably not be affected directly by antifungal prophylaxis. For this reason, studies of anti-infective agents in transplant recipients usually do not use overall mortality as a primary end point. There was no evidence that any deaths in our study were related to itraconazole toxicity.

Breakthrough fungal infections were more common in patients taking prophylactic fluconazole than in those taking itraconazole (14 of 67 patients [20.8%] vs. 4 of 71 patients [5.6%]; P = 0.01). The number of patients receiving empirical amphotericin B (n = 19) and the median duration of empirical amphotericin B therapy (8 days) were identical in the 2 study groups. Empirical amphotericin B therapy was most often used during the neutropenic period immediately after transplantation. Since the decision to initiate therapy with empirical amphotericin B is frequently subjective and the duration of therapy is often too short to adequately treat any established fungal infection, we believe that such therapy should not be used as a primary end point.

Because a unique method was used to administer intravenous itraconazole when this study was started and because a very distinctive taste is associated with oral itraconazole solution, blinding was not done. The number of unrelated transplant recipients enrolled in the study was greater than anticipated, but we agree that randomization could have been stratified by donor type. As discussed in our article, a multivariable analysis using donor type and other relevant factors showed that prophylaxis with itraconazole was still associated with fewer invasive fungal infections. The total cumulative dose of corticosteroids, as well as the mean duration of corticosteroid use, was similar for both study groups.

Antifungal prophylaxis was not given during pretransplantation chemoradiation or after day 100 following transplantation. We do not have data on the incidence of any invasive fungal infections beyond day 180. However, at the University of California, Los Angeles, from December 2001 to July 2003, 53 patients who received allogeneic hematopoietic stem-cell transplants also received itraconazole prophylaxis. These patients were at high risk for Aspergillus infection (median age was 41 years, 79% had advanced disease, and 89% received high-dose corticosteroids for prevention or treatment of graft-versus-host disease), but none developed Aspergillus infection. We note that Dr. Mossad and his colleagues recently reported a similarly low incidence of invasive fungal infections in allogeneic hematopoietic stem-cell transplantation when using an antifungal prophylactic regimen containing itraconazole (5). Thus, we believe that there is sufficient evidence to support the superior efficacy of itraconazole over fluconazole for prevention of Aspergillus and other fungal infections resistant to fluconazole but susceptible to itraconazole.