In their single-center report of antifungal prophylaxis after allogeneic stem cell transplantation (SCT),1 Marr et al refer to our multicenter randomized trial.2 However, the authors failed to note important differences between our study and theirs. In our study,2 we used a dose of oral itraconazole solution (200 mg 2 times daily) that was lower than the dose used by the Seattle group and administered it only after transplantation. Patients unable to tolerate oral medications were switched back to the intravenous form of itraconazole or fluconazole and not re- moved from the study. Despite a lower dose of itraconazole, mean trough plasma concentrations of itraconazole were greater than the 500-ng/mL level targeted for prophylactic efficacy.3 Proven invasive fungal infections occurred in 9% of the itraconazole patients and 25% of the fluconazole patients during the first 180 days after transplantation. The incidence of Aspergillus infection was reduced from 12% in the fluconazole group to 4% in the itraconazole group. Except for more frequent gastrointestinal side effects, itraconazole was well tolerated. The incidences of drug-related hepatotoxicity and renal toxicity were low in our study and much less than those reported by Marr et al. The incidences of hepatotoxicity in both the fluconazole patients and the itraconazole patients reported by the Seattle group were extremely high and much greater than those reported by this group and others in previous studies.3-6 Thus, other factors besides fluconazole or itraconazole were likely respon- sible for liver dysfunction in most of their patients.
On the basis of our results, we introduced long-term itraconazole as routine antifungal prophylaxis in all adult patients undergoing an allogeneic SCT at the University of California, Los Angeles (UCLA). Intravenous itraconazole (200 mg intravenously every 12 hours for 2 days followed by 200 mg intravenously every 24 hours) is started on day 1 after SCT and continued until time of engraftment. After engraftment, patients receive oral itraconazole solution (200 mg every 12 hours) until day 100 after transplantation. After day 100, oral itraconazole is continued in patients still requiring conticosteroids for preven- tion or treatment of graft-versus-host disease (GVHD). Both inpatients and outpatients unable to take oral therapy are returned to intravenous itraconazole. From December 2001 to December 2003, 73 allogeneic SCT patients received itracon- azole prophylaxis. These patients were at high risk for Aspergil- lus infection (median age, 40 years; range, 18-64 years; advanced disease, 78%; previous SCT, 20%; unrelated donor, 41%; high-dose conticosteroids for prevention or treatment of GVHD, 86%; grades II-IV GVHD, 45%). None of the patients developed Aspergillus infection, which is significantly lower than the incidence of Aspergillus infection in similar UCLA allogeneic SCT patients receiving fluconazole prophylaxis be- fore December 2001 (0% vs 13%; P = .004). The only invasive fungal infections were 3 cases of candidemia (2 itraconazole sensitive, 1 itraconazole resistant). Overall survival was 55%, but no deaths were related to fungal infection. Except for nausea and vomiting (19% incidence), itraconazole was well tolerated. There were no cases of hepatotoxicity or renal toxicity attribut- able to itraconazole, and we did not find it necessary to routinely monitor serum drug levels.
In summary, we conclude that Aspergillus can be safely eliminated as a significant pathogen in allogeneic SCT recipients when prophylactic itraconazole is administrated at a tolerable dose and started after the day of transplantation. In view of the difficulty of treating established Aspergillus infections in highly immunosup- pressed patients, we believe that prevention of infection is clearly better than treatment.
Prevention of aspergillosis
We agree with Winston et al's emphasis on prevention; however, we do not agree that the data available definitively demonstrate the superiority of itraconazole. The design of our trial impacted the overall outcomes. Most important was the use of antifungals during conditioning, given the trend to early toxicities in itraconazole recipients who received cyclophosphamide.1,2 This trend appeared to decrease in patients enrolled after the protocol was modified to initiate antifungals after conditioning; however, there were too few patients enrolled to enable comparisons.
We do not have the data necessary to compare toxicities between the 2 trials. Winston et al3 reported the adverse events thought possibly to be related to the study drugs, and we reported the number of patients who developed toxicities independent of cause. More itraconazole recipients developed tripling of baseline total bilirubin (143 of 151, 95%) compared with fluconazole recipients (128 of 143, 86%; P = .02). The incidence of hepatotox- icities in patients who received fluconazole is not different compared with similar patients in our center4 ; the high proportion reflects the definition of “hepatotoxicity.”
Despite the differences in timing and dosing of antifungals in the 2 trials, the outcomes of patients who received itraconazole were remarkably similar. Both studies reported high rates (24%) of gastrointestinal (GI) side effects. Also, the invasive fungal infec- tion (IFI) incidences in itraconazole recipients were not different (13% vs 9%). The only large difference was the incidence of IFI in fluconazole recipients (16% vs 25%, Winston et al study3 ). Outcomes in the Winston et al study were driven by an unusually high number of candidemias (8 of 67 patients, 12%)—much greater than in our study (2.6%) or in prior trials.5,6
While we have learned important lessons, we do not think that the data generated by either trial support the conclusion that itraconazole prophylaxis is a better strategy. In the Winston et al study,3 outcomes may have been impacted by bias between the patients randomized, as more fluconazole recipients received unrelated donor transplants (likewise, increased graft-versus-host disease [GVHD] and therapies). Moreover, the 140 patients enrolled did not provide the power to demonstrate superiority. Finally, it is notable that the apparent IFI protection was not associated with a trend to improved survival. We believe that the conclusion of superiority should be supported at least by a trend toward decreased transplant-related mortality in order to negate the possibility that decreased infection rates are associated with excessive antifungal-related toxicities.
The data presented for the 73 follow-up patients are compelling. Based on the results of Winston et al's trial, we would expect at least 4 IFIs in this cohort. This raises questions: What were the methods of surveillance and the duration of follow-up? Were there other patient or environmental factors that changed? Without having a contemporaneous control, it may be difficult to identify the causative factor(s).
We agree with the assertion that prevention is better than treatment, and we are intrigued by the trends in the Winston et al study3 and the subset of patients enrolled later in our trial. However, we believe that policy decisions should be based on data obtained from properly designed randomized trials rather than trends demonstrated in underpowered studies or retrospective cohort analyses. It is not clear to us that the risk-benefit ratio favors itraconazole for prophylaxis, but we share our colleagues' enthusi- asm that a newer azole may accomplish our common goal of preventing aspergillosis.
Correspondence: Kieren Marr, Program in Infectious Diseases, Fred Hutchinson Cancer Research Center, 1100 Fairview Ave N, D3-100, Seattle, WA 98109.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal