5
RATIONALEThe availability of new antifungal agents with novel mechanisms of action has stimulated renewed interest in combination antifungal therapies. In particular, and despite the limited clinical data, the high mortality of mold infections and the relatively limited efficacy of current agents have produced significant interest in polyene-, extended-spectrum azole-, and echinocandin-based combinations for these difficult-to-treat infections. With the recent publication of the first large randomized trial of antifungal combination therapy to be conducted in two decades (166) and the rapid proliferation of new in vitro and in vivo data on antifungal combinations, we have sought to review the recent work and future challenges in this area.The focus of this review is on the efficacy of antifungal drugs in combination with respect to the extent or rate of killing of the fungal pathogen, although other potential interactions (such as pharmacokinetic drug interactions) can impact efficacy when these agents are used together. The value of giving two drugs because each is separately effective against a group of organisms exhibiting a variety of types of resistance is not specifically discussed, but this also is an obvious and straightforward reason to use a combination of agents.It cannot be simply assumed that the use of two or more effective drugs with different mechanisms of action will produce an improved outcome compared to the results seen with a single agent. Combination antifungal therapy could reduce antifungal killing and clinical efficacy, increase potential for drug interactions and drug toxicities, and carry a much higher cost for antifungal drug expenditures without proven clinical benefit (106). Thus, it is important to critically evaluate the role of combination therapy as new data become available.Conceptual models and terminology. Methods for studying antifungal combinations in vitro and in vivo have differed considerably over time and among investigators. These tools do not differ with respect to their application to combination antibacterial or antiviral therapies and have been discussed extensively and elegantly in the landmark 1995 review by Greco (79). In brief, all approaches to evaluating combinations can be reduced to two elements: (i) a conceptual model for predicting the expected result for a combination and (ii) a set of phrases used to categorize results that are better than expected, worse than expected, or as expected. Although many subtle variations are possible, the underlying mathematical model is based on either the assumption of additive interactions or the assumption of probabilistic (multiplicative) interactions. On the basis of the terminology employed by the author who first carefully described each of these models, the two models can be usefully referred to as the Loewe additivity model and the Bliss independence model (79).The terminology used to place results into interpretive categories is often the subject of debate and confusion. Greco et al. (79) have proposed a set ...