Some respiratory tract infections are associated with asthma development due to inflammatory damage to the airway mucosal and submucosal tissues.¹ Chronic airway inflammation development has mainly been investigated in the context of respiratory viral infections.¹ Fungal exposure is associated with exacerbation of asthma symptoms.² Exposure to airborne bioparticulate matter containing fungal spores (particularly Aspergillus) is a well-established predisposing factor for the development of asthma and other allergic diseases.³ The relationship between asthma and fungal infections like coccidioidomycosis has not been fully described, although one small study linked coccidioidomycosis to worsened asthma outcomes.⁴

Immunocompetent patients with coccidioidomycosis typically present with subacute systemic and pulmonary symptoms. We have hypothesized that Coccidioides could trigger inflammatory airway responses, similar to other fungi, resulting in patient concerns regarding their infection returning, and frequently requiring increased healthcare utilization. To estimate the frequency of new-onset asthma-related symptoms after coccidioidomycosis and identify potentially associated factors, we analyzed data from a large U.S. health insurance claims database.

We used the Merative™ MarketScan® Commercial/Medicare Database (https://www.merative.com/documents/brief/marketscan-explainer-general), which contains health insurance claims data for >48 million patients with employer-sponsored plans including Medicare Supplemental and Medicare Advantage plans throughout the United States during the January 1, 2017 to June 30, 2022 study window. We identified patients diagnosed with coccidioidomycosis using International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) code B38 who had continuous insurance enrollment in the year before and after the date the coccidioidomycosis diagnosis code was first used (index date) during the study window (Figure 1).

To identify incident diagnoses, we excluded patients with a coccidioidomycosis or asthma diagnosis code or an outpatient prescription for a short-acting β₂ agonist prescription (SABA; a standard treatment for asthma) (albuterol, albuterol sulfate, or levalbuterol) in the year before the index date. Patients for whom coccidioidomycosis or asthma was listed on a laboratory or imaging claim alone were not included, to remove instances in which the diagnosis was most likely being ruled out by a healthcare provider.

To compare patients with coccidioidomycosis with vs. without an asthma diagnosis code or a SABA prescription in the year after the index date, we performed bivariable case-case analyses using χ2, Fisher exact, and t-tests at p<0.05. We examined patient demographic characteristics; healthcare provider types visited on the index date; signs, symptoms, and other selected conditions in the 90 days before or on the index date using ICD-10-CM codes (Supplemental Table 1); and outpatient prescriptions. Bivariate comparisons that were significant at p<0.2 were included in a stepwise multivariable logistic regression to estimate adjusted odds ratios (aORs) and 95% confidence intervals (95% CI) for demographic characteristics and clinical features independently associated with asthma diagnosis or SABA prescription.

Last, we identified the proportion of patients who had a new (i.e., not listed in the previous year) asthma diagnosis code or SABA prescription in four separate comparison cohorts meeting the 2-year continuous enrollment criteria. The cohorts were: 1) 12.4 million patients with a healthcare visit for any reason, 2) 1.3 million patients with influenza, 3) 1,412 patients with aspergillosis, 4) 1,320 patients with candidal sepsis during the study window.

MarketScan data are fully de-identified, so this analysis was not subject to review by the Centers for Disease Control and Prevention institutional review board.

Among 6,074 patients with coccidioidomycosis during the study window, 2,748 (45%) met continuous enrollment criteria. In the year before the index date, 296 (11%) were excluded for having a previous coccidioidomycosis diagnosis, 381 (14%) were excluded for having an asthma diagnosis code, and 755 (27%) were excluded for having an outpatient prescription for a SABA, leaving 1,657 patients for analysis. Among those, 208 (13%) had an asthma diagnosis code (n=81) or SABA prescription (n=165) in the year after the index date. The median time from the index date to first asthma diagnosis code or first SABA prescription was 77 days (interquartile range [IQR] 22–242).

The median age of all patients with coccidioidomycosis was 51 years (IQR 39–60). Patients with coccidioidomycosis who had an asthma diagnosis code or SABA were more frequently female than patients who did not (63% vs. 44%, p<0.001) (Table 1). In total, 78% of patients with coccidioidomycosis were from the West, and 95% were from non-rural areas. Patients who had an asthma diagnosis code or SABA were less likely to have visited a family practice or internal medicine provider (30% vs. 38%, p=0.024) and more likely to have visited a pulmonologist (21% vs. 15%, p=0.013) or be hospitalized (22% vs. 12%, p<0.001) on the index date. Overall, disseminated (5%) and cutaneous (3%) coccidioidomycosis were infrequent.

Symptoms and clinical findings that were more common among patients who had an asthma diagnosis code or SABA included: cough (40% vs. 31%, p=0.005), dyspnea (34% vs. 19%, p<0.001), pleural effusion (16% vs. 7%, p<0.001), sepsis (15% vs. 6%, p<0.001), hypoxemia (9% vs. 3%, p<0.001), myalgia (9% vs. 2%, p=0.005), and acute respiratory failure (4% vs. 1%, p=0.002). Other diagnoses more common among patients who had an asthma diagnosis code or SABA included: unspecified pneumonia (34% vs. 25%, p=0.034), smoking (15% vs. 9%, p=0.004), chronic obstructive pulmonary disease or other chronic lower respiratory disease (13% vs. 7%, p=0.006), bacterial pneumonia (10% vs. 4%, p<0.001), and bronchitis or unspecified lower acute respiratory infection (8% vs. 4%, p=0.007). Forty-eight percent of patients with coccidioidomycosis received outpatient antifungal medication, primarily fluconazole (46%), which was more frequently received by patients who had an asthma diagnosis code or SABA compared with patients who did not (55% vs. 45%, p=0.005).

On multivariable analysis, the odds of having an asthma diagnosis code or SABA were highest for women (aOR 2.5, 95% CI 1.8–3.5) and patients with sepsis (aOR 2.3, 95% CI 1.3–3.8), hypoxemia (aOR 2.0, 95% CI 1.0–4.0), or bronchitis (aOR 2.0, 95% CI 1.1–3.7) (Supplemental Table 2).

The proportion of patients with a new asthma diagnosis or SABA prescription in the comparison cohorts was: 6% among those with any healthcare visit, 9% among those with influenza, 11% among those with candidal sepsis, and 20% among those with aspergillosis.

The associations identified in this large study concerning coccidiomycosis and new indicators of asthma-related symptoms advance understanding of relationships between fungi and asthma. Our results suggest that this phenomenon may be more common in patients with more severe pulmonary coccidioidomycosis (as approximated by the need for hospitalization, antifungal treatment, and the presence of clinical complications such as hypoxemia and sepsis). Although our study was not designed to estimate the excess risk for developing asthma-related symptoms attributable to coccidioidomycosis, the proportion of patients with a new asthma diagnosis code or SABA prescription (13%) among those with coccidioidomycosis was higher than that among each of the comparison cohorts except for the aspergillosis cohort. The finding that female patients with coccidioidomycosis were more frequently affected is consistent with the higher national asthma prevalence among women (10%) than men (6%).⁵ We recognize that patients with coccidioidomycosis may have had underlying and previously unrecognized airway hyperresponsiveness or asthma that was quiescent and then unmasked following coccidioidomycosis; however, these data support the hypothesis that Coccidioides may predispose patients to developing inflammation-mediated airway obstruction as observed with other airborne fungi.^2,3

Fungi and other pathogens within the pulmonary bronchioles allow for persistence of antigens that cause direct injury to the airways. For example, Aspergillus proteases drive mucus abnormalities via MUC5AC,⁶ and impaired antigen clearance promotes an allergic response through prolonged stimulation of Th2-type immune cells. Aspergillus proteases and allergens have also been found to cause airway inflammation and remodeling in animal models, and recent work has found an alkaline protease that damages bronchiolar cell junctions triggering inflammation and airway hyperresponsiveness.⁷ Since orthologues to these antigens are present in Coccidioides spp (heat shock protein hsp1, 60S ribosomal protein L3, heat shock protein SSB, fructose 1,6-bisphosphate aldolase among others), a similar pathophysiologic mechanism in humans might explain development of airway hyperresponsiveness following infection.

Other aspects may confound the relationship between coccidioidomycosis and asthma. For example, in some geographic areas, particulate matter levels have been positively associated with coccidioidomycosis cases.⁸ Dust particle inhalation may induce asthma exacerbations by molecular mechanisms involving Toll-like receptors on naïve CD4+ T cells leading to both Th2 and Th17 activation.⁹ Th2 activation leads to airway eosinophilia and mucus hypersecretion, while Th17 activation facilitates neutrophil migration; all this can contribute to asthma-like symptoms and may serve as cofactors enhancing airway hyperreactivity following coccidioidomycosis.

The main limitation of this analysis is potential misclassification inherent in health insurance claims data, which do not necessarily reflect definitive asthma diagnoses using spirometry. The validity of medical coding data for identifying asthma has been studied extensively. A systematic review showed that use of ≥1 asthma diagnosis code over a 1-year period is a commonly used method to identify patients with asthma in administrative data and can be valid.¹⁰ Including asthma medications in the case classification algorithm may improve case detection¹⁰; in one study, nearly half of adults prescribed asthma medication lacked an asthma diagnosis code.¹¹ SABA prescriptions in our study may reflect off-label use to reduce coccidioidomycosis-related symptoms such as wheezing or chest tightness. Our analysis cannot ascertain causality and might also overestimate new-onset asthma in adult patients with a history of childhood asthma or a more distant history of adult asthma >1 year before coccidioidomycosis. Lastly, the dataset lacks information on race/ethnicity; medications administered during hospitalization; and laboratory, imaging, and pulmonary function testing results, all of which may have provided additional insight.

In summary, healthcare providers seeing patients who have recently had coccidioidomycosis should be aware of the possibility for subsequent airway hyperreactivity, particularly among women and patients who have had more severe pulmonary infection. Future investigations using other clinical data sources would advance understanding of this phenomenon.