|Home | About | Journals | Submit | Contact Us | Français|
In a recent report, Schaenman and colleagues (5) describe a case of a Scedosporium apiospermum soft tissue infection in an immunocompromised patient successfully treated with voriconazole. The article focuses on one of the hottest topics in current medical mycology, the emergence of antimycotic-resistant fungal isolates (3). Indeed, Scedosporium apiospermum is, also in our direct experience, one of the emerging fungal pathogens frequently endowed with resistance against drugs used as first-line agents (i.e., amphotericin B and fluconazole) (2). Therefore, we agree to the general message of the paper regarding the need of a prompt and well designed antifungal therapy. However, we would like to address a major point on how this could be achieved. In fact, we disagree that presumptive fungal identification based on aspecific morphological aspects is sufficient to take into account a new drug such as voriconazole as a first-line agent in the management of fungal infections. As admitted by the authors and clearly shown in Fig. 2 of their case report, many fungal genera feature morphological characteristics difficult to discriminate and the identification is not straightforward, especially if specific structures are not usually evident, as may be the case upon direct examination of clinical samples. A clinician making the same assumption as the authors might feel free to treat critically ill patients with voriconazole in the majority of cases, thus putting the whole community at risk for the emergence of new resistances to this valuable drug, as has already and inevitably happened for narrow-spectrum triazoles. Moreover it is still far from being proven that the toxicity profile of the new extended-spectrum triazoles is really safer than that of narrow-spectrum drugs, with severe side effects reported in up to 10% of patients receiving voriconazole (1). We think that a rapid and precise identification, at least at the genus level, is crucial for the prescription of a well designed empirical therapy, but we are convinced that it should be based on objective data. Recently, we addressed the diagnosis of mycotic keratitis using, in parallel with cultural methodologies, a molecular approach based on direct amplification from the biological sample and sequencing, by means of universal fungal primers (4, 6), of genus- and species-specific targets on the fungal genome. In our opinion this molecular approach allowing unequivocal identification of a fungal pathogen, at least at the genus level, in only one day is, together with a more thorough understanding of mechanisms of drug resistance, a real improvement of the conventional mycological diagnosis and represents a correct answer to the clinical questions posed by the availability of multiple classes of antifungal agents.
Mancini and colleagues raise a number of issues regarding the diagnosis and treatment of fungal infections. They suggest, for example, that fungal morphology lacks sufficient specificity to guide antifungal therapy. However, morphological features that lack adequate specificity for precise fungal identification may still provide clinically useful information. For example, Fig. 2 in our report depicts two filamentous fungi (Scedosporium apiospermum and Aspergillus fumigatus) that exhibit a similar morphology of thin septate hyphae branching at acute angles (9). Although no antifungal agent has reliable efficacy against all molds that share this morphology, amphotericin B has historically proven to be ineffective for the treatment of S. apiospermum infection, whereas voriconazole has a spectrum of in vitro antifungal activity that includes both S. apiospermum and Aspergillus spp. (1, 3). For the immunocompromised patient whose case was described in our report, the selection of voriconazole on the basis of this morphology proved to be expedient as the isolated mold was later identified as S. apiospermum.
Despite the usefulness of fungal morphology in some cases, we agree with Mancini and coauthors that improved tests for the early microbiologic diagnosis of invasive mycoses are needed. New assays, however, require proper validation before their introduction into routine clinical use. Mancini and colleagues write that they have modified the broad-range fungal PCR protocol of Pryce et al. (8) for direct amplification of an rRNA gene from clinical samples in the diagnosis of mycotic keratitis. The validation and day-to-day use of fungal PCR, however, is not straightforward. These assays may be particularly prone to false-positive results due to the ubiquity of airborne fungal spores as well as the potential for contaminating fungal DNA in commercial lots of common reagents (5). In addition, a direct amplification approach brings with it issues of DNA extraction efficiency, PCR inhibition, and commensal DNA contamination which may vary with respect to tissue type (4). The substantial scientific rigor required to validate new molecular assays has been documented by national laboratory standard committees (6, 7) and by independent investigators (2, 4). It appears that Mancini et al. have not yet published their study; therefore, we are unable to determine to what degree validation has been addressed.
Even if all validation issues were resolved, the nucleic acid sequencing required by a broad-range PCR approach is not currently performed by most clinical microbiology laboratories. This effectively curtails the present usefulness of broad-range assays for the vast majority of clinicians. In their study, Pryce and colleagues had access to the West Australian Genome Resource Centre for their sequencing needs (8). For those with similar resources, we encourage the continued investigation of molecular assays to facilitate their potential validation for clinical use. For example, we are evaluating in-house broad-range fungal PCR and real-time PCR assays for analysis of clinical samples under a research protocol (D. B. DiGiulio et al., unpublished data).
For the above reasons, among others, we believe that therapeutic decisions informed by the thoughtful interpretation of fungal morphology will continue to be required of clinicians into the foreseeable future. When morphology characterized by thin septate hyphae branching at acute angles is evident (thereby rendering zygomycetes unlikely), we believe that data support the consideration of voriconazole as a first-line agent. As discussed in our paper, toxicity is a consideration when prescribing voriconazole. We believe, however, that voriconazole's toxicity profile must be judged vis-à-vis the toxicities of alternative antifungal therapies. Finally, we share Mancini and coauthors' concerns about the widespread use of newer antifungal agents. In our report, however, we advocate, not the indiscriminate use of voriconazole, but rather a targeted therapeutic approach that is informed by all the evidence available to the clinician, including morphology. When such evidence supports a specific agent as a therapeutic option for the treatment of relatively rare, serious infections, we do not believe that concerns regarding community resistance justify withholding this option from individual patients, particularly those who are critically ill or immunocompromised.