What Causes Fungal Outbreaks?

Public attention has been drawn to recent high-profile outbreaks of mycotic diseases, such as those of fungal meningitis and other infections linked to contaminated steroids [1] and an outbreak of necrotizing cutaneous mucormycosis linked to a tornado [2]. However, fungal outbreaks are more common than most people appreciate, and reports of outbreaks caused by unusual fungal pathogens are increasing. The Mycotic Diseases Branch at the Centers for Disease Control and Prevention (CDC) investigates 3–6 fungal outbreaks per year, many of which are caused by rare fungi with limited diagnostic and treatment options. This is a considerable increase from the 1990s, when the Branch investigated 1–2 hospital-based outbreaks per year, generally caused by yeast and traced to a single source. Although the exact reasons for this increase are unknown, the increased number of patients with impaired immune system may have contributed to this trend.

The majority of fungal outbreaks can be attributed to either environmental exposure or a contaminated product (Table 1). For example, two recent outbreaks were linked to contaminated medications. In 2012, two medications produced by a single compounding pharmacy in Florida were contaminated with Fusarium sp. and Bipolaris sp., respectively, shipped to 15 states, and injected into the eyes of patients undergoing vitrectomies. As a result, 47 patients developed endophthalmitis, and most lost vision [3]. In the 2012 fungal meningitis outbreak, methylprednisolone acetate (MPA) contaminated with Exserohilum rostratum and several other microorganisms was shipped to 23 states, potentially exposing nearly 14,000 individuals to this contaminated medication. As a result, 752 people developed meningitis, arachnoiditis, or spinal/paraspinal abscesses, and 64 patients died, making this the deadliest fungal outbreak to date [1].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Examples of fungal outbreaks.

https://doi.org/10.1371/journal.ppat.1004804.t001

Environmental exposure is the other common cause of fungal outbreaks [2,4]. For example, a recent cluster of Rhizopus delemar infections in a children’s hospital in New Orleans was linked to contaminated linens [4]. In this outbreak, five children died from cutaneous mucormycosis, and R. delemar was isolated from linens, linen shelves, and bins at the hospital [4]. In another example, 13 people developed necrotizing cutaneous mucormycosis caused by Apophysomyces trapeziformis after receiving puncture wounds caused by flying debris during a tornado [2]. Although A. trapeziformis was not recovered from the local environment, whole genome sequence typing (WGST) of the isolates showed that at least three different strains were involved, suggesting environmental exposure [2,5]. In addition, many other outbreaks and clusters caused by dimorphic and other fungal pathogens have been linked to environmental exposure [6–8].

What Role Does Epidemiology Play in Investigating Fungal Outbreaks?

Descriptive epidemiology (i.e., detailed assessment of patients’ demographic characteristics, clinical histories, and the geographic and temporal distribution of cases) is essential for generating hypotheses about the potential source of infection. For example, cases occurring over several weeks to months and scattered geographically suggest a common source outbreak that involves a widely distributed product, especially when case patients have undergone similar medical treatments [1,3]. Conversely, cases occurring among persons with exposure to a common location suggest environmental transmission. Specifically, environmental transmission may be likely among patients with invasive mold infections cared for in the same hospital [4,9,10] or among people with dimorphic fungal infections who participated together in outdoor activities [6,8]. Hypotheses generated through descriptive epidemiology can be tested through analytical epidemiological studies and by microbiological testing of suspected sources (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. CDC epidemiologists are collecting environmental samples in a histoplasmosis outbreak investigation.

https://doi.org/10.1371/journal.ppat.1004804.g001

Why Is Molecular Genotyping Important?

Molecular genotyping complements epidemiological findings. Whereas at least some isolates from common source outbreaks are expected to be genetically identical [1,3], environmental transmission is often associated with multiple strains or species [5,11]. Thus, molecular genotyping is a powerful tool to support or refute epidemiologically generated hypotheses. For example, results of molecular typing directly influenced the response to a cluster of Bipolaris sp. infections detected among patients recovering from cardiothoracic surgeries in three hospitals in Texas, Arkansas, and Florida (United States) that was recently investigated by the CDC. Because of the similarity of the patient population, a common product was suspected but could not be identified. However, multilocus sequence typing (MLST) revealed that none of the patients shared strains with the same genotype and that at least two species were involved, strongly supporting environmental exposure rather than a common source. These molecular results provided critical evidence during an early stage of the investigation and focused investigators on the environment rather than on medications and devices used by these patients.

Similar results were obtained when molecular typing was used to investigate a multistate outbreak of Fusarium keratitis, in which a case-control study implicated a specific type of contact lens solution and suggested a common source outbreak, although Fusarium was not isolated from intact product or the production facility. However, the identification of 19 distinct genotypes of Fusarium from case patients provided evidence for independent contamination events in case patients’ local environments [11].

How Can WGST Help?

When available, conventional methods of molecular genotyping, such as MLST, can rapidly generate results; however, identical patterns obtained by these methods are often difficult to interpret, because they can be attributed to both common origin as well as to the low discriminatory power of the typing method. In particular, conventional genotyping methods are often unable to differentiate among strains from clonal populations with low genetic diversity, such as Cryptococcus neoformans or Cryptococcus gattii [12,13]. For example, in a recent cluster of C. neoformans infections in an Arkansas hospital investigated by the CDC, three of five cases of cryptococcosis were caused by isolates with identical ST5 (A5/M5) genotypes, which is also one of the most common MLST genotypes among environmental and clinical strains [13], making it impossible to determine whether these strains were likely to have been acquired independently or from a common source. In addition, certain genotyping methods, such as microsatellite typing, may generate homoplasy, patterns that look indistinguishable but are not related by descent. Furthermore, for the majority of fungi, conventional molecular genotyping methods are simply not available.

WGST provides a highly sensitive tool for molecular genotyping that can be applied to any pathogen without prior knowledge of the genome, which is especially useful for investigating outbreaks caused by rare fungal pathogens for which population structure information may be unavailable. This method has recently been applied to investigate several fungal outbreaks: (i) to confirm molecular identity of Coccidioides immitis from three organ recipients who shared the same donor [14], (ii) to investigate genetic relationships among isolates of E. rostratum from patients and contaminated methylprednisolone [15], (iii) to confirm genetic identity between isolates of C. immitis from soil in Washington state and a case patient with coccidioidomycosis acquired in that state [16], (iv) to demonstrate multiple origins of the rare mold A. trapeziformis in the tornado-associated cluster of necrotizing cutaneous mucormycosis [5], and (v) to demonstrate genetic identity among strains of Saprochaete clavata, a highly unusual fungal pathogen, associated with a multicenter outbreak in France [17].

The discriminatory power of WGST allows estimation of genetic relatedness among strains of pathogens without prior knowledge of the underlying population structure. For example, WGST analysis of E. rostratum isolates from the fungal meningitis outbreak indicated that 19 isolates from patients and contaminated medication lots had identical genomes of 33.8 Mb and no more than two single nucleotide polymorphisms (SNPs) differentiated any two isolates, confirming a likely single origin of the outbreak strains. By contrast, more than 20,000 SNPs were detected between any two control strains of E. rostratum, confirming the genetic diversity among unrelated strains [1,15]. Conversely, genetically identical isolates of A. trapeziformis as well as isolates separated by thousands of SNPs were identified when WGST was used to investigate the etiology of necrotizing cutaneous mucormycosis, which was consistent with environmentally acquired infections [2,5].

What Are the Limitations of WGST and What Is the Future of Fungal Outbreak Investigations?

Although the results of WGST can significantly enhance epidemiological investigations, this method is still unacceptably slow for most real-time investigations. For example, WGST results for the Exserohilum outbreak were generated 6 months after the initial investigation was completed and therefore provided mostly confirmatory data [15]. In order for WGST to become applicable for real-time investigations, time and cost of generating and analyzing WGST data need to be reduced. WGST methods are already widely used for investigating bacterial and viral outbreaks. However, most fungal genomes are at least ten times larger than those of bacteria and viruses; therefore, considerably more time and resources are needed to generate and process fungal genomes.

Development of a curated public database containing assembled genomes of major fungal pathogens will significantly accelerate analyses and implementation of WGST into public health by providing reference genomes and control strains for assessing genetic diversity in a population as well as facilitate data sharing among institutions. It is also possible that for some fungi, WGST can be substituted with a high-density MLST or a targeted SNP-based typing system that can provide as much information as WGST. As population genomic data for fungal pathogens accumulate, targeted SNPs, MLST loci, or both can be selected for high-resolution targeted genotyping for population genetic studies and molecular epidemiological investigations.

The other significant limitation of current WGST is the difficulty in detecting and identifying fungal DNA directly in human clinical samples against the human DNA background. Proteomic and metagenomic methods have been used for culture-independent detection and typing of viral and bacterial pathogens [18–20], and these methods are being adapted for fungi [21]. Accumulation of fungal genomic data and the development of a WGST database for fungal pathogens will provide a necessary framework for developing metagenomic tools for detection and typing of fungi in clinical samples. In addition, genomic data will facilitate basic research aimed at understanding pathogenesis and improving antifungal therapies. As vulnerable patient populations increase and exposure to pathogenic fungi continues, the number of fungal outbreaks is also likely to increase. We anticipate that novel molecular tools coupled with thorough epidemiological investigation will continue to assume greater importance in recognizing, stopping, understanding, and preventing fungal outbreaks in the future.

Acknowledgments

Disclaimer: The use of product names in this manuscript does not imply their endorsement by the US Department of Health and Human Services. The finding and conclusions in this article are those of the authors and do not necessarily represent the views of the CDC.