Clinical Manifestations and Outcomes of Hospitalized Patients With Blastomycosis: A Provincial Cohort Study

Abstract

Background

Blastomycosis is a rare but potentially life-threatening infection caused by Blastomyces spp that is endemic in the Canadian provinces of Manitoba and northwestern Ontario. Given evolving epidemiology of endemic mycoses, we sought to evaluate the provincial experience managing inpatients with blastomycosis over time. The objectives were to describe the risk factors and predictors of poor outcomes among patients hospitalized with blastomycosis throughout Manitoba.

Methods

We performed a provincial retrospective cohort study by identifying all culture-positive clinical specimens yielding Blastomyces spp within Manitoba, using the Provincial Lab Information System. We included all adult patients who required admission to hospital for infectious manifestations of microbiologically confirmed blastomycosis since 2006. We used a multivariate regression model to predict associations with mortality.

Results

Of 317 patients with culture-positive Blastomyces spp, we included 141 hospitalized patients across 7 hospitals. Overall, in-hospital mortality was 24%. The presence of chest pain at presentation to hospital (odds ratio, 3.24 [95% confidence interval, 1.26–8.83]) was associated with mortality. Hospital admissions, intensive care unit admissions, and mortality have a positive temporal trend over the study period.

Conclusions

Blastomycosis is associated with significant morbidity and mortality, both of which is increased over time. Further studies are needed to evaluate effective diagnostic and treatment strategies to improve outcomes.

Blastomycosis is a rare but life-threatening infection. This study demonstrates that blastomycosis was associated with high morbidity and mortality, which increased during the study period, and that there was a changing geographic distribution of cases during the same period.

Blastomycosis is a fungal disease classically endemic to areas surrounding the Great Lakes of Canada and the United States, and the Ohio and Mississippi River valleys, with evidence of geographic expansion beyond historical areas including throughout the Canadian province of Manitoba [1, 2]. Manitoba is a central Canadian province with a population of 1.5 million people [3]. Additionally, Manitoba occasionally provides health services for people living in northwestern Ontario, where the highest incidence rates of blastomycosis have been reported [1]. Blastomycosis is typically caused by inhalation of spores from the thermally dimorphic fungus Blastomyces spp that appears as mold at ambient temperature and transforms to pathogenic yeast at body temperature [4, 5]. Clinical disease varies in severity from subclinical infection in approximately 50%, to acute or chronic pneumonia and destructive and cavitary pulmonary disease, and may progress to acute respiratory distress syndrome (ARDS) in severe cases [6]. Hematogenous dissemination may occur in 25%–40% of cases involving any organ, but most commonly the skin, bone and joints, genitourinary system, and central nervous system [7–9].

Climate change is thought to contribute to expanding range and changing epidemiology, including new endemic areas of blastomycosis and several other endemic mycoses [2, 10–12]. However, it is uncertain whether the incidence, distribution, and clinical manifestations of blastomycosis is changing in Manitoba, a Canadian province immediately adjacent to the hyperendemic area of northwestern Ontario with the highest incident rates in North America [6]. Previous retrospective reports from Canadian provinces, including Quebec, Ontario, Manitoba, and Saskatchewan, have described the distribution and clinical manifestations of blastomycosis, however, none with contemporary data from the past decade [1, 13–15]. Due to nonspecific presentation as well as lack of rapid diagnostic testing and inconsistent reporting in many areas, epidemiologic surveillance of blastomycosis is needed for enhanced recognition, diagnosis, and treatment [2]. In this retrospective cohort study, we reviewed all hospital admissions for blastomycosis in the province of Manitoba, Canada, to describe the burden and temporal trends of blastomycosis since 2006.

Objectives

The primary objective was to describe provincial temporal trends in hospitalization, disease severity, and mortality of patients hospitalized with blastomycosis. Additionally, we aimed to analyze baseline clinical factors associated with increased mortality. We hypothesized that the burden of disease is increasing, and that geographic expansion in areas with limited prior exposure may be associated with higher rates of hospitalizations, severe disease, and mortality.

METHODS

Study Design

We conducted a retrospective cohort study of all cases of blastomycosis in the province of Manitoba between January 2006 through May 2024. Approval for this study was granted by the Health Research Ethics Board at the University of Manitoba (HS25813; H2023:012).

Population and Inclusion/Exclusion Criteria

Cases of blastomycosis were defined as persons with a clinically compatible presentation and laboratory confirmation of Blastomyces spp from a clinical specimen [16]. All persons who were diagnosed by a clinical culture in a Manitoba laboratory were reviewed. We performed comprehensive manual chart review of first-time admissions on all adult patients (≥18 years) who required an acute admission to hospital for infectious manifestations of blastomycosis. We excluded patients who were admitted electively (eg, for medical procedures) and patients for whom clinical records were unavailable.

Sources of Data, Data Extraction, and Management

The Provincial Lab Information System has recorded all culture-positive specimens with Blastomyces spp throughout Manitoba since 2006. Validated scoring systems were used to capture baseline comorbid status and disease severity, including the Charlson Comorbidity Index (CCI) and National Early Warning Score 2 (NEWS2) [17, 18].

Data were entered and managed using REDCap (version 13.1.37, ©Vanderbilt University). We iteratively created electronic case report forms with a comprehensive data dictionary with input from content and methodologic experts to improve interrater reliability. Variables were defined according to published literature or expert opinion. The database included branching logic and quality checks to ensure data quality. Comprehensive database testing and validation were conducted by multiple users prior to taking the data collection instrument into production.

Key Data Element Definitions

We defined hospital presentation as the date and time of the first hospital visit that led to admission (recorded at triage). We used the earliest documented date and time of a reported result that suspected or confirmed the presence of Blastomyces spp for time of diagnosis, including the results of characteristic findings on microscopy (eg, broad-based budding yeast), histopathology, and/or culture. Duration of symptoms was defined as the number of days prior to hospital admission for which a patient was experiencing any symptom that could be reasonably attributed to blastomycosis (eg, fever). Organ failure was defined as the receipt of high-flow nasal oxygen, noninvasive ventilation, invasive mechanical ventilation (IMV), vasoactive medications, or extracorporeal membrane oxygenation (ECMO).

Statistical Analysis

We performed descriptive statistics including calculating mean ± standard deviation and median (interquartile range [IQR]) for continuous variables, depending on their normality. Categorical variables were summarized as proportions. Continuous variables were compared using the Student t-test or the Wilcoxon signed-rank test, for normal and nonnormal distributions, respectively. Categorical variables were analyzed using χ² analysis. We used Firth regression to predict associations between baseline clinical variables and mortality. Potential variables were selected based on prior literature, clinical relevance, and expert clinical opinion. A backward elimination approach using penalized likelihood ratio tests, with manual adjustments for clinically important variables, was used to select the final model. All statistical analyses utilized a 2-sided α level of .05. Analysis was performed using RStudio version 2024.4.2.764 [19].

RESULTS

Of 317 patients with blastomycosis, 141 adults were hospitalized across 7 hospitals in Winnipeg and surrounding areas (Figure 1). The mean age among hospitalized adults was 46.4 ± 18.3 years (Supplementary Figure 1) and most were male (70%) without prior comorbidity (56%). The mean CCI score was 1.5 ± 1.9, with diabetes (25%) and pulmonary disease (10%) being the most prevalent comorbidities. Thirteen patients (9%) were immunocompromised (Table 1). Immunocompromised conditions included hematologic malignancy (n = 4), solid organ transplant (n = 4), and use of immunosuppressive medications for other medical conditions (n = 5).

Figure 1.

Patient cohort selection flow diagram. Abbreviation: ICU, intensive care unit.

Table 1.

Baseline Characteristics of Patients Admitted to Hospital With Blastomycosis

Baseline Characteristic	Total Cohort (n = 141)	Ward Patients (n = 83)	ICU Patients (n = 58)
Age, y	46.4 ± 18.3	44.8 ± 19.2	48.5 ± 16.9
Female sex	42 (29.8%)	25 (30.1%)	17 (29.3%)
No comorbidities	79 (56.0%)	53 (63.9%)	26 (44.8%)
CCI scorea	1.5 ± 1.9	1.2 ± 1.8	2.0 ± 2.0
Low risk	104 (73.8%)	68 (81.9%)	36 (62.1%)
High risk	37 (26.2%)	15 (18.1%)	22 (37.9%)
Comorbidities
Pulmonary disease	14 (9.9%)	8 (9.6%)	6 (10.3%)
Cardiovascular disease	12 (8.5%)	5 (6.0%)	7 (12.1%)
Immunocompromised	13 (9.2%)	8 (9.6%)	5 (8.6%)
Diabetes	36 (25.5%)	13 (15.7%)	23 (39.7%)
Duration of symptoms, d	29.6 ± 21.8	32.7 ± 22.2	24.9 ± 20.7
Body weight, kg	85.5 ± 22.3	83.4 ± 19.8	87.3 ± 24.2
Active pregnancy	2 (1.4%)	0	2 (3.4%)
Place of residence
Manitoba	110 (78.0%)	64 (77.1%)	46 (79.3%)
Ontario	29 (20.6%)	18 (21.7%)	11 (19.0%)
Nova Scotia	1 (0.8%)	1 (1.2%)	0
No fixed address	1 (0.8%)	0	1 (1.7%)
Travel/reside in northwestern Ontario	65 (46.1%)	42 (50.6%)	23 (39.7%)
Primary site of infection
Pulmonary	113 (80.1%)	61 (73.5%)	52 (89.7%)
Cutaneous	14 (10.0%)	14 (16.9%)	0
Central nervous system	7 (5.0%)	3 (3.6%)	4 (6.9%)
Musculoskeletal	7 (5.0%)	5 (6.0%)	2 (3.4%)
Disseminated	36 (25.5%)	18 (21.7%)	18 (31.0%)
Disease severity
Organ failure <24 h of presentation	16 (11.3%)	0	16 (27.6%)
NEWS2	4.6 ± 4.0	3.1 ± 2.8	6.8 ± 4.4
APACHE II score	…	…	17.4 ± 7.2
CRP, mg/L	153.6 ± 121.5	113.4 ± 105.7	189.8 ± 128.6

Data are reported as No. (%) or mean ± standard deviation.

Abbreviations: APACHE II, Acute Physiology and Chronic Health Evaluation II; CCI, Charlson Comorbidity Index; CRP, C-reactive protein; ICU, intensive care unit; NEWS2, National Early Warning Score 2.

^aLow risk: CCI score 0–2; high risk: CCI score ≥3.

At the time of hospital admission, patients indicated their primary residence in northwestern Ontario (21%) or Manitoba (78%). For residents of Manitoba, 69% were from Winnipeg, 22% were from rural Manitoba, and 9% from northern Manitoba. All reported cases from northern Manitoba occurred in the last 10 years, and the proportion of infections from rural Manitoba increased. Overall, any exposure to the hyperendemic area of northwestern Ontario in the 5 years prior to hospitalization was recorded in 46%. The proportion of cases in patients residing in northwestern Ontario or those with exposure to the area decreased over time, with a larger proportion of cases in patients who resided in rural and northern Manitoba (Figure 2). The annual number of culture-confirmed cases of blastomycosis remained stable over the study period (Figure 3).

Figure 2.

Proportion of cases by geographic region and proportion of endemic exposure over time.

Figure 3.

Epidemiology of blastomycosis and outcomes over time. Abbreviation: ICU, intensive care unit.

Disease Presentation

The mean duration of symptoms was 29.6 ± 21.8 days including cough (71%), dyspnea (51%), chest pain (31%), and hemoptysis (17%). Extrapulmonary symptoms included fever (43%), weight loss (22%), cutaneous (19%), or neurologic (11%) complaints. Most patients (55%) received at least 1 course of antibiotic therapy prior to hospitalization. The median number of antibiotic regimens received prior to admission was 2 with a range of 1 to 5.

The primary site of clinically or microbiologically confirmed infection was in the pulmonary (80%), cutaneous (10%), central nervous system (5%), and musculoskeletal (5%) systems. Disseminated disease (≥2 organ systems involved) occurred in 36 patients (26%) (Table 1).

Treatment and Diagnosis

The median time from presentation to diagnosis and initiation of an antifungal medication was 3.2 days (IQR, 1.2–7.9 days) and 2.3 days (IQR, 1.0–6.1 days), respectively. The most common initial antifungal regimen was with amphotericin B (36% liposomal amphotericin B, 14% amphotericin deoxycholate). Oral antifungals were the initial regimen in 40% of patients, most commonly with itraconazole (Table 2).

Table 2.

Outcomes, Management, and Treatment Complications in Patients Admitted to Hospital

Outcomes	Total Cohort (n = 141)	Ward Patients (n = 83)	ICU Patients (n = 58)
Mortality	34 (24.1%)	3 (3.6%)	31 (53.4%)
Respiratory support
High-flow nasal oxygen	20 (14.2%)	…	20 (34.5%)
Noninvasive ventilation	7 (5.0%)	…	7 (12.1%)
Invasive mechanical ventilation	48 (34.0%)	…	48 (82.8%)
Tracheostomy	11 (7.8%)	…	11 (19.0%)
Cardiovascular support
Pressors or inotropes	43 (30.5%)	…	43 (74.1%)
ECMO	13 (9.2%)	…	13 (22.4%)
Hospital LOSa, d	14 (6–30)	9 (5–17.5)	29 (15.5–49)
ICU LOSa, d	…	…	11.5 (3.25–24.5)
Disease complications
ARDS	35 (24.8%)	0	35 (60.3%)
Pneumothorax	9 (6.4%)	0	9 (15.5%)
Lung necrosis	8 (5.7%)	3 (3.6%)	5 (8.6%)
Time from hospital presentation to diagnosis, db	3.2 (1.2–7.9)	2.7 (1.1–6.7)	4.6 (2.1–9.6)
Time from hospital presentation to antifungal, dc	2.3 (1.0–6.1)	2.0 (0.4–3.6)	4.2 (2.0–7.9)
Initial antifungal regimen
Liposomal amphotericin B	50 (35.5%)	18 (21.7%)	32 (55.2%)
Amphotericin deoxycholate	19 (13.5%)	5 (6.0%)	14 (24.1%)
Itraconazole	52 (36.9%)	46 (55.4%)	6 (10.3%)
Fluconazole	3 (2.1%)	1 (1.2%)	2 (3.4%)
Voriconazole	2 (1.4%)	0 (0.0%)	2 (3.4%)
Unknown	15 (10.6%)	13 (15.7%)	2 (3.4%)
Treatment complications
Acute kidney injury	36 (25.5%)	11 (13.3%)	25 (43.1%)
HAP/VAP	13 (9.2%)	0	13 (22.4%)
Hepatotoxicity	10 (7.1%)	7 (8.4%)	3 (5.2%)

Data are reported as No. (%) or median (interquartile range).

Abbreviations: ARDS, acute respiratory distress syndrome; ECMO, extracorporeal membrane oxygenation; HAP, hospital-acquired pneumonia; ICU, intensive care unit; LOS, length of stay; VAP, ventilator-associated pneumonia.

^aLOS calculated using discharge from hospital or death as the endpoint.

^bData available for 135 patients.

^cData available for 123 patients.

Patient Outcomes

Hospitalizations, intensive care unit (ICU) admissions, and deaths increased over time (Figure 3). ICU admission occurred in 41% (n = 58). None of the patients with cutaneous blastomycosis required ICU admission. In-hospital mortality occurred in 24% (n = 34) (Supplementary Table 1). Of the patients in ICU, 83% (n = 48) were treated with IMV, 19% (n = 11) tracheostomy, and 22% (n = 13) with veno-venous ECMO. Mortality was higher among patients in ICU compared to ward patients (53.4% vs 3.6%), as was the median hospital length of stay (29 vs 9 days) (Table 2). Patients who were immunocompromised had higher mortality compared to those without immunocompromise (38.5% vs 23.4%). Other clinical complications included acute kidney injury (26%), hepatotoxicity (7%), and secondary hospital- or ventilator-associated pneumonia (9%).

Risk Factors Associated With Mortality

In a multivariable model, chest pain at presentation was associated with an increased risk of mortality (odds ratio [OR], 3.24 [95% confidence interval {CI}, 1.26–8.83]), as was higher CCI score (OR, 1.37 per 1-point increase [95% CI, 1.02–1.85]) (Supplementary Table 2). Age and female sex were not significantly associated with increased mortality (OR per year, 1.02 [95% CI, .99–1.05] and OR, 1.14 [95% CI, .46–2.72], respectively). The multivariable model showed good discrimination for mortality with an area under the curve of 0.77 (95% CI, .69–.86).

Ventilation Outcomes

Of the patients admitted to ICU, 83% (n = 48) required mechanical ventilation primarily for hypoxemic respiratory failure (96%). Of the patients requiring IMV, 73% met the 2012 Berlin criteria for ARDS, and the median lowest partial pressure of arterial oxygen to fraction of inspired oxygen ratio in the first 7 days was 57.5 (IQR, 47.25–81.5). Patients with ARDS were ventilated according to established guidelines [20] and many received neuromuscular blockade (86%) and/or prone ventilation (26%) (median, 2 days [IQR, 1–2 days]). Of the ventilated patients, mortality was not significantly higher in patients with ARDS, compared to those ventilated without ARDS (63% vs 54%, P = .814). A total of 11 patients (19%) required tracheostomy, and the average time on IMV was 14.1 ± 8.6 days.

DISCUSSION

This provincial retrospective analysis demonstrated that hospitalized adult cases of blastomycosis have stayed relatively constant in Manitoba over time; however, an increasing proportion of cases are coming from rural and northern Manitoba and areas outside of the usual northwestern Ontario hyperendemic area. Furthermore, blastomycosis-related hospitalization and mortality appear to be increasing over time. Patients admitted to hospital with blastomycosis were typically young, mostly men, and usually without substantive comorbidities. Mortality rates were high in all hospitalized patients (24%) and considerably higher in patients requiring ICU admission (53%) or with ARDS (63%).

The disproportionate sex distribution of blastomycosis is in keeping with previous reports, where men are affected in 58%–85% of cases [1, 13–15]. The increased incidence among men has been hypothesized to arise from potential occupational or outdoor exposures [21], though other biologic mechanisms have not been excluded. Being younger and healthier with few preexisting conditions is also a known risk, with prior reports showing an average age of 38 years and 15%–53% having no comorbidities [1, 15]. Pulmonary manifestations were the primary reason for hospital admission, though we also identified a large proportion of patients with disseminated infection (26%), findings that have been described previously [1, 8, 13–15].

Patients presented with a mean duration of symptoms of 4 weeks prior to admission to hospital. Symptom duration prior to admission was similar between patients admitted to the ward and in ICU (30 vs 28 days). Additionally, patients were frequently (55%) prescribed empiric antibiotics with a median of 2 courses prior to hospitalization. The significant delays in diagnosis, consistent with prior reports, further highlight a need for improved diagnostic testing and a high index of suspicion in the outpatient setting [6, 22]. Due to the nonspecific presentation, clinicians should consider blastomycosis as a possible diagnosis in those who fail to improve with 1 course of antibiotics. The use of novel diagnostic tests (eg, urinary antigen and molecular testing) is an area where further research is needed to assess their impact on time to diagnosis, treatment initiation, and clinical outcomes. After admission to hospital, patients were diagnosed and treated with appropriate antifungals in short order (median of 3 and 2 days following presentation to hospital, respectively), although the ranges were wide, indicating substantial delays (>1 week after hospital admission) in a subset of patients. Delayed diagnosis could be hypothesized to contribute to increased hospitalization and worse patient outcomes; however, this has not been shown in previous studies [22]. The impact of delayed diagnosis is still uncertain and an area that requires further research. The median time to antifungal treatment was shorter than the median time to diagnosis, likely attributed to clinicians experienced in treating blastomycosis starting empiric therapy in those with high-risk and suitable clinical features. Despite this finding, our data showed that many patients presented without a classical risk profile. Patients were increasingly found to reside over a wide geographic distribution throughout Manitoba, most without travel or exposures to the hyperendemic areas of northwestern Ontario. Patients with a primary residence in Ontario decreased (21% vs 41%) compared to prior reports [1].

Mortality from blastomycosis in Manitoba is rising and higher than in earlier provincial reports. In a case series of patients diagnosed with blastomycosis in Manitoba hospitals between 1988 and 1996, mortality was 6.3%, relative to 24.1% in our current inpatient cohort [1]. Provincial-level reporting was not mandatory or possible in this period, possibly indicating differences in cohort selection. Regardless, we observed a 3-fold gradual increase in hospital mortality from 2006 to 2024. This finding may be attributed to a changing geography of Blastomyces spp exposing persons without preformed immunity from prior exposures and leading to more extensive infection. Additionally, patients are increasingly living with greater medical complexity and a higher burden of comorbidities than previously, which may increase the risk for disease severity. One study demonstrated that mortality from invasive fungal infections was increasing and associated with medical complexity such as cancer, immunosuppressive conditions, and other complex medical conditions [23]. While most patients in our cohort did not have comorbidities (56%), we found higher proportions of patients with diabetes (26% vs 11%) and immunosuppression (9% vs 2%) compared to prior reports [1], and higher CCI score was associated with higher mortality. Mortality was higher among immunocompromised patients (38.5% vs 23.4%); the number of cases is small. Further studies are needed to evaluate whether immunocompromising conditions are associated with worse outcomes.

Chest pain was a common presenting feature (31%) and was associated with higher in-hospital mortality (OR, 3.2). While there is some degree of imprecision in this estimate due to small sample sizes, chest pain may indicate infection extension to the parietal pleura, predicting advanced and destructive disease. Closer monitoring of such patients, and potentially more aggressive early management, could be considered. However, further analysis is needed before impacting routine clinical management.

The strengths of this study included the province-wide case reporting of hospitalized patients, size of the cohort, and the comprehensive nature of the data collection. We identified few missing charts, and our chart reviews yielded few missing data points. The data are reflective of a provincial multicenter health system managing patients with blastomycosis in an endemic area.

Our study is limited by the retrospective nature of data collection and the relatively low overall incidence of blastomycosis each year. Our dataset identified patients with culture-positive Blastomyces spp diagnosed within the province of Manitoba; however, patients diagnosed outside of the province but managed within Manitoba, or those without microbiologic diagnosis, could not be identified. Patients with confirmed blastomycosis who did not require an admission to hospital for treatment of blastomycosis were excluded from the study; further studies are needed to demonstrate the full clinical spectrum of blastomycosis outcomes and severity. Speciation and molecular testing of clinical specimens with laboratory-confirmed Blastomyces spp are not routinely done in the provincial laboratory and thus these data are not available for analysis. Changing patterns of referrals for hospitalization of persons from northwestern Ontario to Manitoba hospitals over time may impact the geographic distribution of cases over time. Patient race and ethnic identifiers were not routinely available in the patient records; disproportionate effects of incidence, severity, and mortality in some groups could not be analyzed. Due to the size and retrospective nature of the study, the impact of initiating antifungal therapy on patient outcomes could not be addressed.

CONCLUSIONS

Blastomycosis-related hospitalizations, ICU admissions, and mortality demonstrate a positive temporal trend over time and are associated with substantial morbidity and mortality. While the number of blastomycosis cases in Manitoba stayed constant, the geographic distribution of cases shifted over the study period. Future prospective studies are needed to understand epidemiologic and host factors influencing the risk of infection and poor outcomes.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.