The fungal infections of primary concern in hematopoetic stem cell transplant (HSCT) recipients are candidiasis and aspergillosis. Infections caused by Candida species fall into the distinct syndromes of acute bloodstream infection and chronic infection, or hepatosplenic candidiasis. In both cases, Candida species are primarily acquired from the gastrointestinal tract. Candida parapsilosis is the exception to this rule, as this organism is acquired via the intravenous (IV) catheter route, possibly in association with infected infusates. Aspergillosis is exogenously acquired, via aerosolization. Current debate has focused on the hospital and environmental sources of exposure, as Aspergillus species have been recovered from both air and water supplies.
Risk factors for candidiasis and aspergillosis are summarized in Figure 1. The highest risk for candidiasis is during the neutropenic period, especially in patients who are colonized with the organism during conditioning-related mucositis. The period of greatest risk is followed by an extended period of risk for both candidemia and chronic candidiasis, particularly in association with graft-vs-host disease.
Early risk for aspergillosis is also associated with the neutropenic period; in addition, analysis of cases at the Fred Hutchinson Cancer Research Center has demonstrated an incidence spike in the summer months. However, several centers have documented that the greatest risk period now occurs late after transplantation in allograft recipients, in the setting of acute and chronic graft-vs-host disease. This late risk, particularly when it occurs more than 100 days after transplantation, might constitute the most difficult period to address in terms of disease prevention; generally, this has not been accounted for in prophylaxis trials.
What is optimal antifungal therapy in HSCT patients? For prophylaxis, the characteristics of optimal therapy would include the ability to administer drug for a long duration (especially in allograft recipients), with low toxicities, low cost, ease of delivery, predictable drug interactions, and both anti-Candida and antimold activity. Since we currently lack sensitive and specific diagnostic techniques, optimal empiric therapy should have broad activity against suspected agents. However, increased toxicity and greater cost may be acceptable if such broad activity can be achieved. For treatment of documented infection, the primary requirement is adequate efficacy.
Currently Available Agents
Currently available antifungal agents consist of conventional IV and lipid formulations of amphotericin B(Drug information on amphotericin b), 5-flurocytosine (5-FC [Ancobon]), and the azole drugs, including the imidazoles (clotrimazole [Lotrimin, Mycelex], ketoconazole(Drug information on ketoconazole) [Nizoral], and miconazole) and triazoles (fluconazole [Diflucan] and itraconazole(Drug information on itraconazole) [Sporanox]).
Investigational agents include a liposomal formulation of nystatin(Drug information on nystatin) among the polyenes; the new mold-active azoles voriconazole(Drug information on voriconazole), posaconazole, and ravuconazole; and a number of echinocandins.
Drawbacks associated with amphotericin B treatment include infusion-related and renal toxicities. In addition, there have been reports of resistance among several Aspergillus species, C krusei and C glabrata, and other non-Aspergillus molds. Also, this agent is systemically available only via the IV route, and the lipid formulations are expensive.
The results of several studies agree that the lipid formulations AmBisome and Abelcet cause fewer infusion-related toxicities and nephrotoxicities compared to conventional amphotericin B. Recent debate has focused on whether the decreased toxicities warrant use of expensive products as first-line therapy in either the empiric or therapeutic settings and which product is better.
Among the azole agents, fluconazole(Drug information on fluconazole) is easy to administer (oral and IV forms) and has low toxicity rates and predictable drug interactions. However, it is active only against C albicans. It is currently indicated for use in mucosal candidiasis, prophylaxis, hepatosplenic candidiasis, and candidemia. It is important to note that effectiveness in candidemia has not been demonstrated in randomized trials in compromised hosts, and results of retrospective studies remain controversial.
Itraconazole is now available in an IV formulation and an oral solution form with improved bioavailability, and it is active against both Candida species and molds. Although this drug has been used widely in Europe, it has not been well studied in the HSCT population. Results of several ongoing studies are favorable with regard to absorption, although more information regarding drug interactions is needed. It is currently indicated for maintenance therapy in aspergillosis, but randomized trial data are lacking in the settings of candidemia/candidiasis and antifungal prophylaxis.
Among the investigational agents, voriconazole has been developed in both oral and IV forms and has in vitro activity against molds comparable to that of amphotericin B. Notably, this agent has increased activity against Zygomycetes.
Posaconazole has been developed in an oral form, and may exhibit increased activity against Fusarium species.
Ravuconazole has been developed in oral form and exhibits an extended half-life; it may be associated with less potential for cytochrome P450 system interactions than other azoles. Studies evaluating safety and efficacy of these compounds are underway.
Efforts at preventing disease in HSCT recipients have yielded demonstrable benefits in improving patient outcomes. The ability of prophylaxis or preemptive treatment to reduce morbidity and mortality has perhaps best been documented in the case of cytomegalovirus disease, but use of antifungal prophylaxis also appears to have improved outcomes over the past decade.[6-8] There is, however, a clear need for improvement in preventive efforts.
Although an effective anti-Candida regimen has been established, debate continues regarding optimal dose and duration of drug administration, and no antimold prophylactic treatment has as yet been demonstrated to be effective. Further, although it has been demonstrated that empiric treatment of febrile neutropenia is an important component of prevention, it is also clear that fever is not a sensitive indicator of fungal infection. Better diagnostic methods are needed to improve prevention efforts.
Fluconazole has been shown to be effective in preventing candidiasis after HSCT.[8-9] However, review of major randomized studies indicates that host characteristics affect outcome in an important manner. In a study by Goodman et al, placebo was compared with fluconazole (400 mg daily) until engraftment in a population in which 52% of patients were allograft recipients. Fluconazole was associated with significant reductions in Candida infections and candidiasis-related mortality, but no reduction in overall mortality was observed.
In a subsequent single-center study by Slavin et al placebo was compared with fluconazole (400 mg daily) for 75 days after transplantation in a population that was at higher risk (88% of patients were allograft recipients). Administration of study drug was associated with significant reductions in fungal infections, infection-related mortality, and overall mortality. The findings of this latter study have engendered debate, since the overall mortality benefit could not be statistically explained by the decrease in infection-related mortality. Importantly, this analysis defined fungal infections as those caused by both Candida and Aspergillus species, the latter of which fluconazole has no activity against.
We have subsequently performed an analysis of 8-year survival in this patient population, finding that fluconazole was associated with significant long-term survival benefit in allograft recipients (17% reduction; P = .0018), and the benefit is associated with prevention of candidiasis and related mortality. These findings emphasize both that fluconazole is effective prophylaxis for candidiasis and that preventive efforts and treatment trials must distinguish between patient types and types of fungal infection.
Rates and distribution of Candida infections in HSCT patients at Fred Hutchinson Cancer Research Center (FHCRC) prior to (1980 to 1986) and after (1994 to 1997) the introduction of fluconazole prophylaxis are shown in Figure 2.
The use of fluconazole prophylaxis was associated with a decrease in incidence of candidiasis from > 10% to 4.6% during these periods. As can be seen, the frequency of C albicans infection has been dramatically reduced, whereas there has been some emergence of infection due to C glabrata and C krusei. Importantly, the increase in C parapsilosis infection is associated with an outbreak of infected intravenous infusate during 1994, not with the use of fluconazole. Concomitant with the dramatic decrease in candidiasis, aspergillosis has become the predominant infection over the past decade. The 1-year cumulative incidence of disease has steadily increased at FHCRC since 1993, with current rates exceeding 10% in allograft recipients (data not published).
As noted, the optimal dosage of fluconazole (400 vs 200 mg) necessary for prophylaxis has not been defined. However, no dose studies are likely to be performed in this setting, due to the warranted focus on investigation of treatments to address aspergillosis in HSCT patients. A number of studies of aspergillosis prophylaxis have been performed; however, these studies have involved patient populations that have an underlying low incidence of aspergillosis, weakening statistical analyses with inadequate sample sizes.
In a study of twice-daily aerosolized amphotericin B vs a standard amphotericin B regimen in a patient population including low-risk leukemic patients, no significant difference in rates of aspergillosis were observed (4% vs 7%) and rates of amphotericin B toxicity were high.
In a study of itraconazole vs placebo in 405 neutropenic patients, itraconazole was associated with a lower rate of candidiasis, but equivalent rates of aspergillosis. In a study of itraconazole vs fluconazole in 445 neutropenic patients, itraconazole was associated with lower rates of both candidiasis and aspergillosis. However, the small number of proven aspergillosis cases in each of these studies impairs our ability to evaluate efficacy of prevention.
Randomized trials evaluating prophylaxis with both itraconazole and posaconazole in higher-risk HSCT patients are ongoing. In addition, a study of the echinocandin FK463 in allogeneic and autologous HSCT patients has been initiated.
Lipid Amphotericin B
Comparative trials in febrile neutropenia have included a study of liposomal amphotericin B (AmBisome) vs conventional amphotericin B in 338 patients in which the liposomal formulation (1 and 3 mg/kg) was associated with a reduced incidence of nephrotoxicity (10% to 12% vs 24%, respectively). In another comparison of liposomal with conventional amphotericin B, liposomal amphotericin B was found to be associated with reduced toxicity and a decrease in breakthrough infections. However, the observation that most breakthrough infections were caused by azole-susceptible species (C albicans or C tropicalis) raises a concern about whether the difference in breakthrough rates is associated with differences in type and effectiveness of antifungal prophylaxis between the two arms.
In a study comparing AmBisome at 3 and 5 mg/kg with the amphotericin B lipid complex Abelcet, AmBisome at 5 mg/kg was associated with decreased rates of infusion-related toxicity and nephrotoxicity. 
Ongoing studies include a comparison of liposomal amphotericin B and voriconazole and a comparison of liposomal amphotericin B and the echinocandin caspofungin(Drug information on caspofungin). Although decreased toxicities associated with the lipid products have led to increasing use over conventional amphotericin B, it must be realized that one potential shortcoming of the studies performed to date is that overall efficacy is difficult to assess with only a short (1 to 2 week) duration of follow-up.
Fewer randomized trials have been performed to evaluate treatment of confirmed infection. The results of one study, which evaluated outcome of suspected and proven infections in patients who were randomized to receive either liposomal amphotericin B (5 mg/kg) or conventional amphotericin B (1.0 mg/kg/d), suggest that efficacy of treatment might be greater with the lipid-based product. These studies generally have shown a decrease in toxicity with liposomal products, but efficacy continues to be debated; it is important to note that interpretation of efficacy outcomes in these treatment trials is also made difficult by inclusion of mixed patient populations. One ongoing trial is comparing conventional amphotericin B and voriconazole in HSCT patients, and more are needed.