We combined deidentified data on blastomycosis cases reported in Arkansas during January 1995–May 2018, Louisiana during January 1987–October 2018, Michigan during January 2007–December 2017, Minnesota during January 1999–December 2018, and Wisconsin during January 1990–December 2017. We also used the Louisiana Hospital Inpatient Discharge Database to identify additional cases among hospitalized patients in Louisiana during 1999–2014.

We included data elements that were collected by >3 states. We considered event date as the earliest date associated with the case; for example, symptom onset, or first healthcare visit, laboratory test order, or public health report. We considered all laboratory tests recorded as positive for blastomycosis to be positive, even without an explicitly stated qualitative or quantitative result. Negative blastomycosis test results were not routinely available; therefore, we did not include these in the analysis.

We used patients’ state and county of residence to calculate annual state-specific incidence and county-level mean annual incidence per 100,000 persons by using yearly population estimates from the US Census Bureau, Population Division, Vintage 2015 Special Tabulation (https://www.census.gov). We used χ², Fisher exact, and t-tests to identify factors independently associated with hospitalization or death, the Cochran-Armitage test for trends in the proportion of patients who were hospitalized or died, and negative binomial regression to assess incidence trends, and we considered p<0.05 statistically significant. We also compared demographic features and outcomes among cases associated with outbreaks (outbreak cases) and those not associated with outbreaks (nonoutbreak cases) for Minnesota and Wisconsin, the 2 states that reported outbreaks during the surveillance periods we examined. Human subjects review by the Centers for Disease Control and Prevention determined this project to be consistent with nonresearch public health surveillance.

We summarize blastomycosis surveillance data from 5 states and provide a broad update on the basic epidemiology of this enigmatic and underrecognized disease. Many patients experienced severe outcomes and diagnostic delays. Our results show that blastomycosis is underdetected, even in states where it is reportable, and that more standardized and in-depth surveillance, ideally in additional states, would help public health professionals better identify highest-risk groups and emerging areas for targeted prevention messaging.

Blastomycosis often results in severe illness, even in previously healthy persons (9), but this observation might be influenced by underdetection of asymptomatic or milder, self-resolving disease. The high hospitalization rate of 57% noted in this analysis demonstrates that blastomycosis surveillance detects severe cases, which is typical for passive disease surveillance. We found an annual mean of <200 cases/year; a hospitalization rate of 57% suggests that ≈110 patients are hospitalized each year from states where blastomycosis is reportable. In contrast, ≈1,000 blastomycosis-associated hospitalizations occur nationwide (10,11), showing that the limited surveillance likely underdetects cases nationally.

The average time of >1 month from symptom onset to diagnosis indicates delays in seeking healthcare, delays in diagnosis, or both. This time interval is consistent with a previous report describing a median of 23 days between examination at a healthcare facility and a median of 2.5 courses of antibacterial medications for presumed bacterial infection before pulmonary blastomycosis was correctly diagnosed (3). Earlier diagnosis might reduce unnecessary antibacterial drug use, time, and resources invested in searching for alternative diagnoses and could potentially improve patient outcomes. Therefore, greater public and provider education about blastomycosis is needed, especially in areas where blastomycosis is less commonly recognized.

The high proportion of patients with positive confirmatory laboratory tests, such as culture and microscopy, likely reflects detection of more severe cases because serologic tests for blastomycosis offer only presumptive evidence of infection (12), and serologic tests were not included in most states’ case definitions (Appendix). The associations between older age and confirmatory test types with hospitalization point to severe illness, and are unsurprising; however, why women were more likely to be hospitalized is unclear but could be related to delayed diagnosis or underdiagnosis of less severe disease in women. More blastomycosis hospitalizations typically occur among men (10,13), although a recent study found female sex was independently associated with death in blastomycosis patients with acute respiratory distress syndrome (14). The increased risk for hospitalization among persons of non-white races lends further evidence to the existence of blastomycosis-related health disparities, as previously suspected (15–17). Further studies could help determine whether these differences are related to genetic predisposition (18), involvement in outdoor activities resulting in exposures to Blastomyces, or access to medical care (19).

Reliance on often invasive and time-consuming tests such as culture and microscopy for diagnosis likely is a key factor in underdiagnosis of blastomycosis because these tests might not be ordered until tests for other diseases have been negative. Accordingly, most specimen types in our analysis were from bronchoalveolar lavage and lung and other tissue, which likely required biopsy. Given the prolonged time to diagnosis we identified, improved noninvasive diagnostic methods with high sensitivity and specificity for blastomycosis are needed for earlier and more frequent testing, which could prevent hospitalizations and deaths.

Consistent with previous reports, Wisconsin had the highest number of cases and incidence of the 5 states where blastomycosis is reportable, with mean annual incidence in several northern counties >20 cases/100,000 population. Peaks in incidence in Wisconsin corresponded to a known outbreak at a yard waste site in 2006 (20), an outbreak likely associated with multiple sources in 2010 (21), and an outbreak linked to recreational tubing on the Little Wolf River in 2015 (22). For case-patients in these outbreaks, younger age and higher likelihood of being non-White was consistent with our findings (20,21). In addition, the finding that patients with outbreak-associated cases had less severe outcomes could reflect detection of milder cases through enhanced case detection efforts during outbreak investigations. However, outbreaks comprised <6% of cases overall, suggesting that most cases occur sporadically, which also is true for histoplasmosis and coccidioidomycosis. Of note, most of northern Wisconsin is rich in soils classified as spodosols, which are characterized by high concentrations of organic matter in coarse, often sandy, particles (23). Blastomyces spp. are thought to dwell primarily in organic-rich soils. However, spodosols also occur widely in northern Michigan, where disease incidence was not elevated, and are less common in northern Minnesota, where incidence was higher. Further study, including the role of soil types, could elucidate the natural habitat of these fungi.

For most states in this analysis, the relatively stable incidence and hospitalization rates over time were consistent with a previous analysis of blastomycosis-related hospitalizations during 2000–2011 (10). Another study found a decline in blastomycosis-associated deaths nationwide during 1990–2010 (15); the reasons for the increase in deaths we observed during 2007–2017 are unclear but could reflect improvements in case follow-up, a decline in reporting of less severe cases, or other surveillance changes over time.

The limitations of our study include that pooling surveillance data based on different blastomycosis case definitions is fundamentally problematic; however, few other data sources would enable analyses of thousands of cases, which is helpful for studying this uncommon disease. Furthermore, some states’ case definitions changed over time. Although blastomycosis was reportable in each state during the years included in this analysis, Arkansas did not have a formal case definition, and Michigan did not have one until 2012. Wisconsin classified all cases as confirmed until September 2015, when their case definition changed to include confirmed and probable case classifications; for outbreaks in Wisconsin, a positive serologic blastomycosis test plus an epidemiologic link was sufficient to be considered a case. Moving forward, the standardized blastomycosis case definition from the Council of State and Territorial Epidemiologists will enable more robust comparisons between states and stratification of confirmed and probable cases.

Combining data from different times in each state is an additional potential limitation. Some states’ surveillance systems underwent changes during the analysis period; for example, data elements were added or removed, resulting in inconsistent denominators in the pooled analysis. For certain variables, such as race and ethnicity, missing data or values of “unknown” were common and demonstrate that information can be challenging to obtain because substantial time and resources often are needed to conduct case investigations (24). Data about environmental exposures, immunocompromised status, body site of infection, occupation, illness duration, and treatment were not available consistently from every state. Wisconsin and Minnesota conducted extensive follow-up on cases (19,25), providing deeper insight into state-specific features of blastomycosis. Collecting these types of data in a standardized way in additional states could help identify high-risk populations and activities and help inform prevention efforts.

In summary, blastomycosis remains a rarely reported but severe disease in most areas where it is under public health surveillance. Our findings indicate that blastomycosis likely is underdetected. Blastomycosis also can occur in areas outside those where it is commonly recognized (4) and might be emerging in new areas, such as east-central New York (5). Surveillance for blastomycosis in more areas and collection of more standardized, detailed data could help identify emerging geographic hotspots or clusters, new risk factors, and other epidemiologic patterns. Increased awareness among healthcare providers and the public could lead to faster diagnosis and treatment for blastomycosis patients.