The concept of combination therapy for the treatment of infectious diseases is certainly not new. Combinations of drugs have been administered as single formulations (eg, trimethoprimsulfamethoxazole) and as multidrug regimens for many years. Combination antifungal therapy has the potential advantage of improving outcomes of infection by increasing microbial killing, providing different tissue distributions, and allowing for dose adjustments of drugs that have concentration- related toxicities. However, with few exceptions, the utility of combination antifungal therapy has not yet been well documented in large clinical trials. Because systemic fungal infections are most often observed in patients who are immunocompromised, the term "combination antifungal therapy" could be loosely interpreted to mean a combination of an antifungal drug with an immunomodulating agent. However, in its strict sense, combination antifungal therapy refers to the combination of two or more antifungal drugs, which is the focus of this article. Background Fungi may be unicellular organisms (eg, yeasts) or multicellular filamentous organisms (eg, molds), generally characterized by the presence of a chitin and glucan-containing cell wall and an ergosterolcontaining cell membrane. Each of these structures may be the site of action of a particular class of antifungal agent. The polyenes (amphotericin B, lipid formulations of amphotericin B(Drug information on amphotericin b), and nystatin(Drug information on nystatin)) bind to ergosterol in the cell membrane, thereby destabilizing the membrane and causing leakage of cellular components. The azoles (fluconazole, ketoconazole(Drug information on ketoconazole), itraconazole(Drug information on itraconazole), and voriconazole(Drug information on voriconazole)) act within the cell to inhibit ergosterol synthesis so that the cell membrane produced during cell division is inadequate. The antimetabolite flucytosine(Drug information on flucytosine), or 5-fluorocytosine, is a water-soluble, fluorinated pyrimidine analog. Once taken up by the microbial cell, 5-FC is converted to 5-fluorouracil, which inhibits the synthesis of both DNA and RNA. The echinocandins are members of a new class of compounds that inhibit the synthesis of an essential polysaccharide-1,3,beta-D-glucan- in the cell wall of select fungi. There are potential theoretical advantages and disadvantages to using two or more antifungal drugs. Potential Advantages
The potential advantages of combining antifungal agents are as follows:

• An increase in antimicrobial activity secondary to an increase in both the rate and the degree of microbial killing.
• A decrease in microbial drug resistance.
• An increase in the spectrum of activity, particularly applicable to infections in which the exact identity of the infecting fungus is unknown.
• Enhancement in the tissue distribution of the two drugs, an important consideration in certain disseminated CNS infections.
• Reduction in drug-related toxicity, particularly if the dosage of a drug that has dose-related toxicities can be reduced.

Potential Disadvantages
The potential disadvantages of combination therapy include the following:

• Antagonism between the two drugs, so the activity of one or both drugs is reduced.
• An increase in the potential for drug-related toxicity.
• Increased risk of drug-drug interactions.
• Increased cost compared with single-drug therapy.

Antagonism between two antimicrobial drugs may occur because of either an interference with the mechanism of activity or an adverse effect on drug metabolism. Either of these events would adversely affect the available concentration of drug, so that microbial resistance might be induced. Combination Antifungal Therapy Before any consideration is given to combining antifungal drugs, it is imperative that the mechanisms of action and the metabolism of each antifungal agent be understood. Our knowledge of combination therapy is limited; however, we do have the benefit of experience with different agents used in different infections and, in most cases, we know how the individual agents attack fungi. Amphotericin B and Flucytosine for Cryptococcal Infections
In patients with cryptococcal infections, the standard of care is the use of flucytosine plus amphotericin B. It is well known that flucytosine should not be used alone because microbial resistance against this agent develops relatively frequently. Also, flucytosine has dose-limiting toxicities, particularly targeting the bone marrow. Early in vitro studies of Cryptococcus neoformans showed that when flucytosine was combined with amphotericin B, the effect was at least additive and possibly synergistic. One interesting clinical study in non-HIV-infected patients who had C neoformans infection demonstrated that a relatively high dose of flucytosine (150 mg/kg) in combination with a relatively low dose of amphotericin B (0.3 mg/kg), given for 6 weeks, was associated with a faster response and a higher cure rate than those achieved with amphotericin B (0.4 mg/kg) given alone for a longer period.[1] This study set the stage for later studies of cryptococcal infections in patients with AIDS. The first study in patients with HIV infection was a retrospective analysis. Surprisingly, it reported that flucytosine, at doses varying from 75 to 100 mg/kg, plus amphotericin B did not affect survival but was associated with increased drug toxicity.[2] However, the findings from a small prospective trial suggested that the outcomes were better with a combination of flucytosine (at a dose of 150 mg/kg) and amphotericin B (at a dose of 0.7 mg/kg).[3] No definite conclusions can be drawn from these trials, but the conflicting messages generated by the retrospective study and the small, prospective clinical trial supported the performance of a definitive clinical trial examining one- vs two-drug therapy for the treatment of cryptococcal meningitis in patients with AIDS. Four hundred patients with AIDS and cryptococcal infection were randomized to receive either amphotericin B (0.7 mg/kg) plus flucytosine (100 mg/kg) or amphotericin B alone for 2 weeks.[4] After 2 weeks, patients who were stable or who had shown improvement were further randomized to receive either high-dose itraconazole (400 mg/d) or high-dose fluconazole(Drug information on fluconazole) (400 mg/d) as "consolidation therapy." At the end of 2 weeks, no statistically significant differences in mortality rates or symptoms were noted with the combination therapy compared with monotherapy.[4] However, there was a trend toward a better composite success rate with combination therapy: 60% of the patients in the combination-therapy group had cerebrospinal fluid (CSF) culture results that were negative for C neoformans, compared with 51% of the patients in the monotherapy group (P = .06). Also, the mean opening CSF pressure on lumbar puncture was lower in the patients who had negative cultures. Both a negative culture result and lower CSF opening pressure are associated with better outcomes in patients with cryptococcal meningitis, and, for this reason, the combination of amphotericin B and flucytosine is now considered first-line therapy for these patients. The consolidation phase of the study showed that fluconazole was associated with a higher rate of sterile CSF than was itraconazole, but the difference was not statistically significant.[ 4] The various studies and trials in patients with cryptococcal infections have reinforced the concept that retrospective analyses are inherently flawed because of treatment bias and the many uncontrolled variables that are not accounted for. Randomized trials have indicated that the clinical response to a combination of drugs may depend on both the doses used and the duration of use. Also, surrogate end points must be evaluated in studies that may not be adequately powered to demonstrate differences in survival. In the cryptococcosis studies, better outcomes were associated with the higher dose of amphotericin B and the lower dose of flucytosine, given for a shorter duration than in previous studies. Most likely, the success of therapy depends on both microbial killing and host toxicities. Amphotericin B and Azoles in Candida Infections
The use of amphotericin B with an azole antifungal agent is somewhat controversial.[5-7] Because the polyenes, such as amphotericin B, bind to ergosterol in the cell membrane and the azoles inhibit the synthesis of ergosterol, it had been proposed that combining the two types of drugs would be less effective than monotherapy. Both in vitro and animal studies have yielded conflicting results as to whether amphotericin B and an azole are actually antagonistic.[5,6,8,9] The outcome with this combination seems to depend on the particular azole used, the sequence of administration, and the animal model used. Sugar et al[5] found that combination therapy with fluconazole and amphotericin B was not antagonistic in a murine model of invasive candidiasis. In contrast, Louie et al[7] found evidence of antagonism. Another study found that itraconazole and amphotericin B were antagonistic.[ 6] In a randomized, blinded clinical trial that compared high-dose fluconazole (800 mg/d) with or without amphotericin B (0.7 mg/kg/d) in non-neutropenic patients with at least one blood culture result positive for Candida species, combination therapy was associated with a lower rate of persistently positive blood culture results.[ 10] The success rate was slightly higher in patients treated with combination therapy than in patients receiving fluconazole alone (Figure 1). This study did not detect antagonism between fluconazole and amphotericin B; however, the combination was more nephrotoxic than fluconazole used alone. In this study, the difference between fluconazole used alone and fluconazole in combination with amphotericin B was related to the APACHE II score.[10] Among patients with the lower APACHE II scores and those with the higher APACHE II scores, no difference between the response rates with combination therapy and monotherapy was noted. However, a difference was clearly evident among patients who had intermediate APACHE II scores, with those who received the combination therapy having higher response rates. These observations were interpreted to mean that the more intensive combination therapy would not significantly alter outcome for patients with mild disease or for those with the most severe disease-essentially, in those two populations, host factors were driving clinical outcomes.[10] The fact that antagonism was not demonstrated is a very important finding. However, this combination is unlikely to become a standard approach, given the increase in toxicities and the availability of a new class of drugs (the echinocandins) that have activity against Candida species. Amphotericin B and Azoles in Aspergillosis
Based on in vitro and animal data, the combination of a polyene and an azole for aspergillosis is not well justified, except in the setting of diagnostic uncertainty. Several studies using various combinations of azoles and amphotericin B in different animal models of infection have shown antagonism between these types of antifungal agents, and at least two studies have failed to show an effect (Table 1).[11-15] Nevertheless, because of a pressing need to treat invasive mold infections, this combination has been used at times, largely forced by clinical desperation. Combination Therapy With Newer Agents: The Echinocandins
As noted previously, the echinocandins, of which caspofungin(Drug information on caspofungin) is the first, inhibit synthesis of an essential polysaccharide in the cell wall, as opposed to having an effect on the ergosterol in the plasma membrane. Hence, antagonism may not be an issue. Several in vitro studies of caspofungin and amphotericin B used against Candida species and Aspergillus fumigatus used a checkerboard design.[16] With this design, the fractional inhibitory concentration index (FICI) is calculated by dividing the minimal inhibitory concentration (MIC) of the combination of agents by the MIC of each of the drugs. An FICI less than 0.5 indicates synergism of the combination, and an FICI greater than 4 indicates antagonism between the two antifungal agents. In one study that evaluated the combination of amphotericin B and caspofungin for A fumigatus, an FICI of 0.66 was documented, suggesting that the combination was not antagonistic, possibly even approaching synergism. Other in vitro studies have explored a combination of various echinocandins combined with polyenes or azoles against Aspergillus (Table 2). One in vitro study showed that the activity of a combination of the echinocandin micafungin(Drug information on micafungin) and liposomal amphotericin B was either additive or indifferent (ie, neither antagonistic nor synergistic).[17] However, in a murine model of invasive pulmonary aspergillosis, a combination of micafungin and amphotericin B was synergistic.[18] In another in vitro study, a combination of caspofungin and amphotericin B had additive to synergistic effects.[19] The combination of caspofungin and voriconazole was found to be synergistic in both an in vitro study[20] and a guinea pig model.[21] In an in vitro study of the activity of a combination of voriconazole and caspofungin against several different species of Aspergillus (Aspergillus fumigatus, Aspergillus terreus, Aspergillus flavus, and Aspergillus niger), no antagonism was demonstrated, and the combination was additive in 42%, synergistic in 46%, and indifferent in 12% of isolates tested. The combination of caspofungin and voriconazole resulted in a reduction in the geometric mean MIC of each drug-from more than 64 to 16 μg/mL for caspofungin and from 1 to 0.25 μg/mL for voriconazole.[ 20] In a study of experimentally induced invasive aspergillosis in immunosuppressed guinea pigs, treatment included caspofungin (1 or 2.5 mg/kg/d) with or without oral voriconazole (5 mg/kg/d), voriconazole alone, or amphotericin B (1.25 mg/ kg/d).[21] Death occurred in 12 of 12 untreated controls, compared with 4 of 12 and 6 of 12 of the animals that received caspofungin at 1 mg/kg/d and 2.5 mg/kg/d, respectively, and 3 of 12 of those treated with amphotericin B. No deaths occurred in the animals given the combination of caspofungin and voriconazole or those given voriconazole alone. Given that the rates of survival with the combinations of voriconazole and caspofungin were no different from that with voriconazole alone, the question naturally arises as to the value of adding caspofungin to voriconazole. In this study, the advantage of the combination was demonstrated by a reduction in colony counts in tissues of animals treated with the combination of drugs (Figure 2).[21] In addition to the above-described in vitro and animal studies, a number of small, retrospective clinical reports have evaluated the efficacy of combination antifungal therapy involving the echinocandins. An early study of caspofungin in patients with pulmonary aspergillosis was reported from the Memorial Sloan-Kettering Cancer Center in New York.[22] In this retrospective study, the case records of 30 patients who had acute leukemia complicated by possible or proven pulmonary aspergillosis were reviewed. All patients had been treated with a combination of caspofungin and either amphotericin B or liposomal amphotericin B. Sixty percent of patients had a favorable response to the addition of caspofungin, and 20% of patients who had a favorable response had complete resolution of pulmonary aspergillosis. One of the limitations of this study was that only a minority of the patients had proven invasive aspergillosis. A second limitation was that caspofungin was added to ongoing therapy with amphotericin B at a median of 12 days, which leads to the question of whether the timing of combination therapy affects the outcome. Finally, no comparator group was used. Another study of 48 patients with either documented invasive aspergillosis (23 patients) or probable invasive aspergillosis (25 patients) was reported. In this cohort, the majority of patients received a combination of caspofungin and liposomal amphotericin B as salvage therapy because of a failure of amphotericin B monotherapy during the previous 7 or more days.[23] All patients were immunosuppressed: 50% had leukemia, and the remainder were allogeneic transplant recipients; 33% of the patients had high APACHE II scores (16 or higher). The overall response rate was 42%. As with the previously described study, one of the limitations of this study is that the diagnosis was uncertain in a substantial number of patients, and no comparator data were available. These two studies suggest that large medical centers that routinely treat immunosuppressed patients are using these combinations quite aggressively, despite the current lack of controlled data. Whether the combination of the drugs has an advantage that outweighs the potential toxicities or provides substantial benefit is currently unknown. What can be said about the echinocandins is that the in vitro studies, the animal models, and the early clinical studies have provided very promising results that support further exploration in randomized clinical trials to evaluate their use in aspergillosis. However, a randomized trial is a very complicated endeavor requiring careful consideration of the drugs, doses, patient population, and end points. Such a trial would have to enroll a large number of patients. For instance, to show a 10% improvement with the combination of voriconazole and caspofungin compared with voriconazole alone, a minimum of 570 evaluable patients would have to be enrolled; however, given the severity of the underlying disease in these patients, a higher sample size may have to be initially used. Conclusions

• Cryptococcosis-For cryptococcal infections, the combination of amphotericin B and a low dose of flucytosine is standard therapy, based on results of a good randomized trial.
• Candidiasis-For the treatment of candidiasis, a combination of amphotericin B and an azole probably provides more effective clearance of the organism from the bloodstream than does fluconazole alone. One can debate whether this combination is optimal therapy in light of the availability of echinocandin antifungals and the increased toxicities observed.
• Aspergillosis-For aspergillosis, convincing data as to the best initial treatment are lacking. For patients who are being treated with one agent and whose disease is progressing, it appears reasonable to add a second agent, such as an echinocandin, or to switch classes of antifungals. However, using combination therapy as primary therapy is an untested hypothesis. Since drug-related toxicities may negate the potential antimicrobial benefits, a randomized trial for primary therapy most certainly needs to be performed.