US database study: burden and healthcare resource utilization in adults with systemic endemic mycoses and aspergillosis

A total of 34,751,020 patients were queried across the 30 US sites. The annual prevalence of SM was relatively constant: 2012 (n = 6120), 2013 (n = 6080), 2014 (n = 5430), 2015 (n = 5550), 2016 (n = 5600) and the projected data for 2017 (n = 5265). The most prominent fungal infections were histoplasmosis (43.8%), aspergillosis (37.7%) and coccidioidomycosis (14.0%) (Table 1). In total, data from 33,230 patients with SM were available for initial analysis (Table 1); 39% showed treatment with antifungal medications in their EMRs (n = 13,181). Geographically, patients received care in 30 sites, with three sites accounting for over half of all cases: University of Texas Southwestern Medical Center (23%), University of Iowa Hospitals and Clinics (15%) and University of Arizona (11%) (Figure 1.)

Patients with specified SM (histoplasmosis, aspergillosis and blastomycosis) with at least one prescription for one of 11 specific antifungal agents were further categorized: otherwise healthy cohort (n = 2321; 17.2%), aged ≥65 years (n = 769; 5.7%), immunosuppressant medication (n = 5513; 40.7%), cancer (n = 3284; 24.3%), transplant recipients (n = 1099; 8.1%) and HIV/AIDS (n = 532; 3.9%) (Table 2). Patient characteristics for the SM study cohorts are provided in Table 2. The mean ages across the study cohorts were consistent, (50–57 years), excluding the elderly cohort. The HIV/AIDS cohort was notable for three characteristics, youngest mean age (50 years), mainly male (highest proportion, 79%) and lowest proportion of white race (44%). Patients with SM were far more likely to be immunocompromised (77.0%) (cancer [24.3%], transplant recipients [8.1%], HIV/AIDS [3.9%] or receiving systemic immunosuppressant therapy [40.7%]) than not immunocompromised (22.9%) (otherwise healthy [17.2%] or age >65 [5.7%]).

Table 3 compares symptom burden in SM patients who were aged over 65 years (elderly) or immunocompromised versus the otherwise healthy patients (reference cohort) with systemic mycoses. The symptom burden experienced by elderly SM patients for almost all symptoms and symptom classes was broadly similar to the reference cohort, exceptions were the higher prevalence of malaise and fatigue, edema and metabolic symptoms (hypertension, hyperlipidemia and diabetes mellitus) in elderly patients.

All metabolic symptoms, including hypokalemia, were significantly more prevalent in patients on immunosuppressants. Among patients on immunosuppressants, symptom burden was significantly higher among those with SM compared with the control cohort, but the difference in symptom burden for the SM versus control cohort was even more pronounced in the cohorts with cancer, transplant recipients and patients with a history of HIV infection. In those three cohorts, virtually all symptoms were significantly more prevalent in the SM cohort than in the reference cohort, with the exceptions of diabetes mellitus in the HIV/AIDS cohort and abdominal/pelvic pain and acute upper respiratory tract infection in the immunosuppressant cohort (Table 3).

Symptom burden in immunocompromised patients with SM (on immunosuppressants, with cancer, transplant recipients or HIVAIDS) was also compared with patients with the same immunocompromising feature or condition, but without systemic mycoses (control cohorts). The control cohorts (matched non-systemic mycoses) had substantially larger samples sizes: immunosuppressants (n = 4,166,383), cancer (n = 774,313), transplant recipients (n = 37,913) and HIV/AIDS patients (n = 91,275). Patient cohorts with SM and cancer or HIV/AIDS show a uniformly consistent higher prevalence for all symptoms listed (Table 4). Furthermore, when compared with control cohorts of comparable patients without SM (measured as Δ%), SM patients with cancer or HIV/AIDS had much higher rates of constitutional symptoms (cancer 999%; HIV/AIDS 760%), respiratory symptoms (1190%; 624%), metabolic symptoms (630%; 450%), gastrointestinal symptoms (844%; 576%) and infections (i.e., influenza and pneumonia, acute upper respiratory infection and severe sepsis) (2265%; 955%) (Table 4).

Patients with SM and on immunosuppressants also bore a significantly higher symptom burden compared with their control cohort, albeit with numerically lower percentage differences compared with the cancer or HIV/AIDS patients. Transplant recipients appeared to show the lowest percentage differences in symptom prevalence compared with their respective control cohort, notably, in pain and gastroesophageal reflux disease, the differences were not significant. For three metabolic symptoms, rates were significantly lower in systemic mycoses patients compared with control cohort [hypertension (-16%), hyperlipidemia (-27%) and diabetes mellitus (-16%).

Antifungal treatment utilization among the patients with SM is summarized for the six study cohorts in Table 5. Drug utilization adds up to over 100% because some patients received more than one antifungal agent, either in combination or sequentially, during the data collection period. Fluconazole was the most commonly prescribed treatment when measured across all cohorts, with 56.0% (weighted average) of all patients using this medication, followed by voriconazole (53.5%) and itraconazole (19.5%). Liposomal amphotericin B was prescribed at an intermediate rate (11.6%) (Table 5). Treatment utilization was consistent for the top three antifungals across cohorts; however, more itraconazole was prescribed on a percentage basis than voriconazole for HIV/AIDS patients. Slightly more voriconazole than fluconazole was prescribed for patients on immunosuppressant medication (45.70 vs 41.09%) and patients with cancer (69.86 vs 66.86%). Notably, although itraconazole use was third-most frequent antifungal across all cohorts, its use in transplant recipient patients was lower and it was ranked 7th as antifungal treatment in these patients (Table 5).

Healthcare resource utilization (HCRU) over a 6-month period was assessed and patients with SM consistently utilized all three HCRU modes (ER [emergency room], hospitalisation or outpatient visits) more frequently than their corresponding control cohorts, with the exception of ER visits for patients on immunosuppressant medications, 20.2% of patients with SM compared with 32.4% of control patients on immunosuppressants presenting to ER (Table 6). Highest hospitalization rates were observed in SM transplant recipients (74.5%), followed by non-SM transplant recipients (65.1%). SM patients who underwent transplant and SM with cancer showed the highest rates of outpatient visits, 82.3 and 67.0%, respectively (Table 6). Generally (with one exception noted above), across all patient cohorts with SM, there was higher HCRU compared with the control study cohorts without systemic mycoses, these differences were statistically significant (p < 0.01) (Table 6).

This study evaluated outcomes associated with the burden of illness for patients with SM, focusing on those eligible for itraconazole treatment according to its approved uses in the USA. In this study, a majority of patients were immunocompromised in some manner (77.0%), with the data showing that these patients (whether cancer or HIV patients, those on immunosuppressant therapy or transplant recipients) have a significantly higher overall burden of illness than SM patients that are otherwise healthy or elderly (≥65 years old). Given the magnitude in the frequency of reported symptoms for the immunocompromised cohorts, these patients have a substantially greater burden of illness arising from systemic mycoses. Because the database captures symptoms regardless of cause, it is also possible that these patients had a greater symptom burden due to comorbid conditions and/or adverse effects of treatment, in addition to the effects of SM infection.

The relative percentage of constitutional symptoms which were experienced in both otherwise healthy and elderly SM cohorts may be secondary to comorbidities and age-related physiological changes leading to increased patient symptomatology (Flevari et al.) [13]. This is consistent with our study findings where a higher incidence of constitutional symptoms was observed in the elderly healthy cohort compared with otherwise healthy younger (i.e., < 65 years old) patients. It would be useful to have additional detail beyond what was available in this database in order to better understand the relationship between age and burden of systemic mycoses. It was also noted that SM patients in the immunosuppressed cohort experienced higher rates of constitutional symptoms when compared with both the otherwise healthy SM cohort and the immunosuppressed control groups that did not have SM, findings that suggest the immunosuppressed population may have had a greater burden of disease.

The largest differences in symptom burden in SM immunocompromised patients (i.e., immunosuppressant medication, cancer, transplant or HIV) compared with otherwise healthy SM patients were observed in cancer patients as shown in Table 3. Additionally, in cancer patients with SM compared with cancer without SM, there was commonly 10–20× greater incidence of many reported symptoms, across all symptom classes. As these patients are frequently neutropenic secondary to recent chemotherapy it is not surprising that more severe symptoms were observed in this cohort. This high symptom burden is also likely exacerbated by the documented concomitant mental, physical and emotional distress that accompanies an active cancer diagnosis along with symptom burden for patients with systemic mycoses [14].

Respiratory symptoms were common in patients with SM in this study, reflecting that the predominant site of infection in these patients is the lung [15,16]. The rates of pulmonary symptoms in all immunosuppressed patients with systemic mycoses were higher than the otherwise healthy counterparts that have systemic mycoses. Constitutional symptoms such as fever, malaise, pain, abnormal weight loss, were increasingly prevalent across the spectrum with SM from otherwise healthy patients (4–11%); elderly patients (5–16%) to cancer patients (22–57%) (Table 3). Other sites of infection for the target organisms include the bloodstream, splenic and liver abscess, urinary tract and central nervous systemic sites [17].

The observed increases in prevalence of metabolic symptoms in SM patients compared with control cohorts may be due to direct effects of infection on metabolism or exacerbation of comorbidities during active and ongoing infection. Alternatively, infection of endocrine organs, thus adversely affecting regulatory controls on metabolism may contribute to observed rates of metabolic symptoms [18,19]. In addition, current antifungal agents (polyenes, azoles) used for prophylaxis and/or treatment of mycoses also have adverse endocrine and metabolic effects, including hypoadrenalism, hypogonadism, hypoglycemia, dyslipidemia and hyperkalemia or hypokalemia [18,20,21].

Gastrointestinal infection by SM pathogens is uncommon [22]. Gastrointestinal symptoms tend to be nonspecific (e.g., nausea) unless distinct organs such as the liver, bowel, or pancreas are directly invaded. The results reported in this study for the rates of gastrointestinal symptoms arising from SM are consistent with similar findings from prior research, with the highest incidences in the cancer cohort, relatively less of an increase compared with control in the HIV cohort and the transplant cohort mostly resembling the otherwise healthy cohort.

With rising incidence of invasive fungal infections, the burden of symptoms in patients at high-risk of fungal infections is a high priority for healthcare professional and decision-makers. In one large epidemiological study of invasive fungal infections utilizing data from an integrated US health network (Intermountain Healthcare), invasive fungal infections were associated with a high rate of hospital admissions (76.2%) and high 1-year all-cause mortality rate of 28.8% [2]. The conditions or diseases that place patients at higher risk for invasive fungal infections are themselves associated with high symptom burden. This study documented the very high rates of constitutional, respiratory, metabolic, gastrointestinal and infection-related symptoms that are attributable to SM or the accompanying treatment. This study also documented substantially increased HCRU rates attributable to SM.

Given the high burden of SM in terms of symptoms and HCRU, there is a clear impetus for timely diagnosis of SM with identification of the pathogen, which can improve outcomes through prompt initiation of the appropriate antifungal therapy. Currently, many hospitals must send testing off-site for aspergillus and histoplasmosis, which can delay diagnosis and initiation of targeted treatment [23,24]. Of note, 56% of prescriptions in this study were for fluconazole, which is inactive against aspergillosis (38% of diagnoses) and less effective than itraconazole against histoplasmosis and blastomycosis, which together accounted for 47% of diagnoses [8,24,25]. Thus, a substantial number of patients were receiving fluconazole for a disease not optimally treated with this agent. It is possible that fluconazole is sometimes prescribed empirically until fungal testing results are available or fluconazole may be chosen due to limited drug–drug interactions and predictable bioavailability [26].

Future research in larger populations, including among patients with other forms of SM beyond the three pathogens studied here, would be valuable in establishing the burden of SM more broadly. This study was intentionally narrowly focused on pathogens of interest based on our objective of understanding the patient population eligible for itraconazole treatment in the USA, but there would be value in conducting similar analyses on other systemic fungal infections, including candidiasis, cryptococcosis, mucormycosis and pneumocystis infection. It would also be useful to study treatment patterns and outcomes of treatment using real-world evidence, to better understand whether patients are being adequately treated or if there is unmet clinical need in SM.

The main limitation in utilizing ‘live’ cross-sectional EMR data relates to the lack of longitudinal data, patients diagnosed with systemic mycoses based on ICD-10 codes were the starting point for data capturing. Capturing EMR data at a fixed point in time precludes re-running the analysis to include or exclude features not accounted for at the time of querying. It is important to note that the EMR database interrogated gives insight into the rate of symptoms experienced per cohort, but not the severity, nor the duration of impact at the patient level or clinical significance of these symptoms (potentially responsible for delays in next chemotherapy cycle etc). In addition, it is acknowledged that there are limitations to the estimates of healthcare resource utilization, as the total HCRU cannot be fully determined from the EMR data available (e.g., supplemental nursing support, at-home care, rehabilitation support, etc.). HCRU was measured as the percentage of total number of patients in the specified group that sought care. This would not account for repeat visits; hence, HCRU would likely be underestimated if there were repeat visits by patients in those groups (i.e., SM patients) shown to use more healthcare resources. HCRU was also not captured at a high level of granularity; for example, hospitalization is not further divided into intensive care unit versus standard unit care. Finally, although different lipid formulations of amphotericin B are available, this level of granularity was not available using the data extraction tools from this EMR network.

To the authors’ knowledge, this is the first study which evaluates the real-world symptom burden and HCRU attributable to systemic mycoses in patients with additional select comorbid conditions. Of note, the observation that the majority of patients who contract a systemic fungal infection are immunocompromised is significant. Given the rising age profile of the US population, associated increases in the rates of serious illnesses including cancers, conditions requiring organ transplants and other chronic conductions that result in long-term compromised immune responses, it is expected that there will be a corresponding increase in high-risk patients vulnerable to invasive fungal infections. The nature of the comorbidities affecting these patients causes substantial symptomatic burden of illness.

This study provides a basis and reference point for further investigation of quality of life in patients with SM, as well as corresponding healthcare resource utilization and economic burden.

Parts of this study were previously published at the International Society for Human and Animal Mycology, Congress of the European Hematology Association and International Immunocompromised Host Society Symposium in 2018.

G Lewis and R Jordan are employees of Mayne Pharma, which sponsored the research. A Forsythe is an employee of Purple Squirrel Economics, which acted as a consultant to Mayne Pharma. GR Thompson III acted as a consultant to Mayne Pharma. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

K Sandman of Purple Squirrel Economics provided editorial support. This work was sponsored by Mayne Pharma.