Assessment of Dermatophytosis Treatment Studies: Interpreting the Data
October 2015 | Volume 14 | Issue 10 | Supplement | s48 | Copyright © 2015
Theodore Rosen MD
Department of Dermatology, Baylor College of Medicine, Houston, TX
and doing a quick comparison of cure rates? The unequivocal answer to this rhetorical question is a resounding “no.” Why is this so? Official product information sheets enumerate the results of clinical trials. Such trials demonstrate, to the satisfaction of the Food and Drug Administration (FDA), that an agent is safe and “effective.” However, there are marked differences between trials. Limitations inherent to published studies, including those submitted as pivotal trials, include most prominently differences in study design, data collection, and data analysis. Such differences make direct comparison of efficacy across clinical trials nearly impossible, and even preclude reliable systematic reviews and meta-analysis.
The purpose of this article is to remind the HCP of the various parameters to consider when attempting to assess the data on any of the approved agents for dermatomycoses. While emphasizing recently approved agents, a critical approach to antifungal studies is widely applicable: to new agents as well as old, oral as well as topical, and across all types of dermatophyte infection.
When and Where Study Conducted
There is a rather remarkable disparity regarding which dermatophytes are predominant in differing geographic regions of the globe; and the exact etiologic organisms have indeed undergone dramatic and significant changes during the twentieth century.3,7 Several well-known examples of this phenomenon include the replacement of Microsporum audouinii by Trichophyton tonsurans as the leading cause of tinea capitis in the United States, as well as the replacement of Epidermophyton floccosum by T. rubrum as the leading cause of tinea cruris.
Thus, it is important to consider when and where antifungal studies were performed in order to assess the data. Studies done many years ago may not reflect currently dominant etiologic fungi and, even if comparative in nature, will not include the newer agents enumerated previously. Moreover, as the location of antifungal studies will determine which species are predominant, the results may or may not be relevant to a given clinician. For example, the results of a clinical trial involving tinea capitis done in Iran, Libya, Palestine, Spain, or Sweden would be nearly irrelevant for an American HCP because Tricholosporum violaceum is quite common in those nations but vanishingly rare in the U.S.7,14
Finally, one other point is intuitively obvious. The location of a study may greatly influence the ethnicity of the participants. It may not be possible for an American practitioner, serving a community of Caucasian, Afro-American, Asian, and Hispanic individuals, to extrapolate antifungal treatment decisions based upon the results of a clinical trial done in Iceland, a land of remarkable homogeneity (Norse and Celtic ancestry). There is certainly medical evidence that some drugs have significantly different pharmacokinetic and pharmacodynamic properties in varying racial and ethnic groups.15 To the extremely limited extent that such knowledge exists relating to antifungal agents, a HCP must consider trial subjects’ race and ethnicity when assessing clinical trial data and its relevance to his/her own practice.
There are a few critical parameters related to how a study was actually conducted which may impact reported data. These include whether the study was open label or double-blinded, prospective or retrospective, single-center or multi-center, and compared with placebo or another established active agent.
Many initial proof-of-concept and some subsequent trials are open label, especially in Phase 2 investigations. It was admirably demonstrated that, at least for onychomycosis, the efficacy rates of open label studies are substantially higher compared with randomized controlled trials, and may therefore overestimate actual cure rates.16 This is likely due to bias engendered when both subjects and investigators “believe” that the study agent is effective (or an open label study would not be performed in the first place). A single center study is more likely to have bias (either favorable or unfavorable) than when multiple, geographically diverse centers are involved.
Finally, it has been clearly and repeatedly demonstrated that all approved antifungal drugs are superior to a placebo (including vehicle) control.17 While this is the traditional standard for conducting antifungal studies, and is still an acceptable manner in which to obtain FDA approval, going forward most impactful antifungal studies should be done as a direct comparison between 2 active agents. In the absence of well-performed and adequately powered head-to-head treatment investigation, the real difference in efficacy between 2 agents designed to treat the same dermatomycosis may not be assessable.17 Even when head-to-head studies ostensibly compare the same drugs for the same disease, there may be quite different results.18-20 Such differences may relate to study design features, inherent study flaws, and the manner of efficacy data collection and reporting (discussed below).
Inclusion and Exclusion Criteria
A critical appraisal of antifungal trial data would certainly start with ascertaining who was (and who was not) allowed to enroll,