Cryptococcus spp. are basidiomycetous yeast, with two species, C. gattii and C. neoformans, causing nearly all human cryptococcal infections [1]–[3]. C. neoformans typically causes disease in immunocompromised individuals [4], [5], and is an important and common cause of opportunistic infections in HIV/AIDS patients worldwide [6]–[8]. In contrast, C. gattii has historically been characterized as a rare pathogen, with disease confined to tropical and subtropical climates, particularly the highly endemic regions of Australia and Papua New Guinea [9], [10]. Until recently, C. gattii was thought to affect primarily immunocompetent persons living in these regions [11]–[15].

Since 2004, an outbreak of C. gattii infections has been documented in the United States Pacific Northwest states of Oregon and Washington [16]. The outbreak in these states is thought to have originated in, and spread from, British Columbia, Canada, where infections caused by the same C. gattii genetic types have been documented since 1999 [17], [18]. Genetic sequencing has demonstrated that C. gattii can be divided into four molecular types, denoted as VGI, VGII, VGIII and VGIV [19]; these molecular types can be distinguished by various genetic techniques and have different geographic distributions [20]–[23]. The emergence of C. gattii infections in Oregon, Washington State, and British Columbia is primarily due to the clonal expansion of three genetic subtypes belonging to the molecular type VGII, called VGIIa, VGIIb, and VGIIc [24]–[26]; these have been referred to as ‘outbreak-strain’ subtypes. Outbreak-strain subtypes are different from those found in historically endemic Australia and Papua New Guinea (and other areas of the world), where infections are most frequently caused by nonclonal strains of C. gattii, most commonly VGI [11], [27]–[29]. Clinical differences between C. gattii infections in the United States Pacific Northwest and historically endemic areas have been described. While C. gattii in historically endemic areas has been reported to infect primarily immunocompetent persons, causing meningoencephalitis [11], [27], [28], [30]–[32], C. gattii infections in Oregon and Washington State occur frequently in immunocompromised persons and present most often as respiratory illness [33].

Current guidelines for the management and treatment of cryptococcal disease from the Infectious Disease Society of America (IDSA) recommend antifungal treatment that varies by site and severity of infection [34]. The recommended initial treatment (RIT) for severe pulmonary disease, central nervous system (CNS) disease, and other disseminated disease (such as cryptococcemia) is amphotericin B (AMB) and 5-flucytosine (5FC); for non-severe pulmonary disease, the RIT is fluconazole, with itraconazole and posaconazole as acceptable second-line agents [34]. These recommendations are unchanged from previous IDSA guidelines for cryptococcosis, released in 2000, which were available when the majority of patients in this report were diagnosed [35].

Although IDSA guidelines for treatment of cryptococcal disease are based primarily on data from C. neoformans infections in HIV and solid organ transplant patients, these guidelines are intended to apply to patients with C. neoformans or C. gattii infections. A limited number of C. gattii-specific recommendations were included for the first time in the 2010 IDSA guidelines and are based on data from C. gattii infections in historically endemic areas, the only data available at the time of the guideline-writing [34]. These recommendations pertain mainly to patients with cryptococcomas, which previous data have suggested are more common in patients infected with C. gattii than C. neoformans [31], and include consideration of surgery for patients with large cryptococcomas, increased radiologic and follow-up evaluations for those with cryptococcomas or hydrocephalus, and possible use of AMB/5FC in patients with large and/or multiple pulmonary cryptococcomas [34].

Given the genotypic and clinical differences between C. gattii infections in the United States Pacific Northwest and cryptococcal infections due to either C. gattii from historically endemic areas or C. neoformans, applicability of the current IDSA guidelines to C. gattii patients in Oregon and Washington State is unknown. We conducted a retrospective cohort study of C. gattii infections reported in these states to evaluate the relationship between IDSA guideline-recommended initial antifungal treatment and clinical outcomes.

We describe the initial antifungal treatments utilized in United States Pacific Northwest C. gattii infections and subsequent patient outcomes. Patients in this analysis, as previously described [16], [33], were frequently immunocompromised or had serious comorbid conditions and most commonly presented with pulmonary disease. The overall case-fatality rate for this cohort was high (19%). We found that while a substantial minority of patients did not receive IDSA guideline-recommended initial therapy, the receipt of alternative initial treatments was not equally distributed across all C. gattii infections. Fewer patients with pulmonary infections compared with central nervous system infections received IDSA guideline-recommended initial therapy. Among patients with isolated pulmonary infections, fewer with severe pulmonary infections received recommended initial therapy compared with those persons with non-severe infections. Among the patients who received alternate initial treatment, most were ‘under-treated’, either through failure to receive 5-flucytosine with amphotericin B (for patients with CNS, bloodstream and severe pulmonary infections), or failure to receive any treatment (for patients with non-severe pulmonary infections). Receipt of an alternative initial therapy was associated with a non-significant trend towards increased mortality (approximately 30% excess mortality) in the three months after diagnosis, particularly among patients with pulmonary infections.

There are a number of reasons why IDSA-recommended initial therapy might not have been used with patients in this cohort. While infectious disease clinicians are likely to be aware of the IDSA guidelines for cryptococcal disease, many patients are initially treated by clinicians without formal infectious disease training who may not be aware of the IDSA guidelines. Specifically, they may not be aware that severe pulmonary cryptococcosis should be treated in the same way as central nervous system cryptococcosis, leading to under-treatment of patients with severe pulmonary infections. Additionally, as C. gattii infections in the United States Pacific Northwest appear to be clinically different from C. gattii infections in other areas of the world, some clinicians who are aware of the IDSA guidelines in Oregon and Washington State may initially deviate from IDSA-recommended therapy due to concerns about the generalizability of the guidelines to their patients with C. gattii infection. Finally, clinicians may not have used guideline-recommended initial therapy due to matters beyond their control, such as patient contraindications to medications, insurance restrictions, or drug shortages. While we were unable to evaluate why clinicians chose, in a minority of patients, to pursue alternative treatments, our data suggests that there might be some benefit in adhering to IDSA guideline-recommended initial treatment in United States Pacific Northwest C. gattii patients, particularly those with pulmonary disease. Further research into the reasons for use of alternative initial treatment regimens is needed.

Pulmonary cryptococcosis presents a number of clinical challenges in diagnosis and treatment. Unlike cryptococcal meningitis, a common HIV-related opportunistic infection, pulmonary cryptococcosis is much less-commonly-recognized clinical entity, even among HIV-infected persons [36]. Diagnostic delays for patients with pulmonary cryptococcosis, as seen in this cohort, have been documented previously [37]–[40]. In terms of treatment, while IDSA guidelines do specify use of antifungal treatment for pulmonary cryptococcal infections – even mild disease - and the use of amphotericin B and 5-flucytosine in ‘severe’ pulmonary disease [34], [35], the quality of the evidence for both recommendations is limited and based on ‘…opinions of respected authorities…clinical experience, descriptive studies, or reports of expert committees’ [34]. Unlike for cryptococcal meningitis [41]–[44], no randomized controlled trials evaluating best treatments for pulmonary cryptococcosis have been published, and divergent opinions exist in the literature on the utility and optimal type of antifungal treatment for these patients. Some clinicians have suggested that asymptomatic or minimally symptomatic pulmonary cryptococcosis in immunocompetent persons requires no antifungal treatment at all [45], [46], while others have recommended azole drugs or amphotericin B in all cases [47], [48]. In Australia, where guidelines recommend amphotericin B and 5-flucytosine for all but mild/asymptomatic pulmonary cryptococcosis, Chen et al recently published outcomes data on ten patients with isolated pulmonary C. gattii infection. The majority of patients were treated with amphotericin B and 5-flucytosine and only one death was reported [49], raising the question of whether a more aggressive approach might be warranted among patients with pulmonary C. gattii infections. Larger-scale evaluations of patients with pulmonary cryptococcal infections, including patients with C. gattii infections from both previously-recognized endemic areas and the United States Pacific Northwest, are needed to identify the most appropriate treatment and improve outcomes.

We chose to evaluate initial antifungal treatment, and not treatment later in the course of disease, for several reasons. First, initial treatment, termed induction therapy, for cryptococcal disease has been shown to have a strong impact on mortality. Studies of HIV-infected patients during the early years of the HIV epidemic demonstrated that induction therapy for cryptococcal meningitis with fluconazole resulted in worse outcomes than induction with amphotericin B and 5-flucytosine [46]. Similarly, treatment with amphotericin B alone (i.e., without 5-flucytosine) has been shown to be inferior to combination amphotericin B and 5-flucytosine therapy for induction [44], [45]. In addition, we believed that selection of the initial antifungal drug is less likely than subsequent treatment choices to be influenced by outside factors (e.g., patient contraindications, adverse reactions) and therefore more indicative of physician preference and knowledge. However, it is possible that consolidation and maintenance drug choices as well as duration of therapy also may influence patient outcomes. Unfortunately, we were not able to evaluate that relationship in this study.

This analysis included several limitations. First, this patient group included those with C. gattii infections that were severe enough to required hospitalization; thus, our findings regarding treatment and outcomes are likely not applicable to mild, self-limited pulmonary C. gattii infections. However, few of these mild infections have been identified in this cohort and it is unclear how frequently they occur. Second, due to the retrospective nature of this study, not all patients received identical diagnostic testing (e.g., lumbar puncture, chest imaging, bronchoalveolar lavage); this may have led to incomplete ascertainment of all sites of infection. Third, these results are specific to patients with C. gattii infection in the United States Pacific Northwest, and may not be generalizable to patients with C. gattii infection in other areas. Finally, the number of patients in our evaluation was small, particularly in subgroup analyses. More data, ideally from prospective studies or clinical trials, is needed to understand the relationship, if any, between site of infection, initial antifungal treatment, and outcomes in this population.

This is the first evaluation of the effect of initial antifungal treatment on patient outcomes in the North American outbreak of C. gattii. We show that a substantial minority of patients are not getting the current guideline-recommended initial antifungal therapy, which may be associated with improved outcomes. Timely diagnosis and appropriate treatment for patients with C. gattii infection will continue to be a clinical question as the emergence of C. gattii in Oregon, Washington State, British Columbia, and elsewhere continues. C. gattii infections in persons with and without recent travel history to the United States Pacific Northwest or British Columbia are increasingly being reported throughout the United States [14], [29], [33], [50], [51]; this increased visibility will raise more questions about the best treatment for patients with C. gattii. As our identification of C. gattii infections improves and diagnoses increase, as they are likely to do, careful collection of treatment-related data from patients with these infections will be vital to improving outcomes.