Epidemiology of Cryptococcosis Among Patients With Commercial Health Insurance and Patients With Medicaid, United States, 2016–2022

Abstract

In 2 large health insurance claims databases, cryptococcosis prevalence was 3.4 cases per 100 000 commercially insured patients and 6.5 per 100 000 Medicaid patients. Prevalence was higher among males, non-Hispanic Black patients, and residents of the Southern United States, likely reflecting the disproportionate burden of HIV in these populations.

Cryptococcosis is a severe fungal disease caused by the environmental pathogen Cryptococcus. C. neoformans and C. gattii are the most common illness-causing species. Although both species can cause various clinical manifestations, C. neoformans is often associated with central nervous system infection [1]. Cryptococcosis typically affects immunosuppressed people, particularly those with advanced HIV infection, and over 150 000 cases of cryptococcal meningitis and 112 000 deaths occur worldwide in people with HIV [2].

In the United States, the proportion of cryptococcosis cases among people with HIV has decreased in the last 2 decades, while the proportion of cases in transplant recipients, people with other immunosuppressing conditions, and seemingly immunocompetent people has increased [3–5]. Population-based surveillance previously estimated an incidence of 0.4–1.3 cases per 100 000 population in 2 metropolitan areas in 2000 [6]. However, current cryptococcosis incidence in the general population is unknown. More recent data focus on hospitalizations (∼4900 nationwide in 2019) [7], which may overrepresent the most severe cases. Additionally, previous multistate studies have been somewhat geographically limited [4, 5, 8]. To provide an update on the epidemiology of cryptococcosis in the United States, we analyzed 2 large health insurance claims databases to estimate disease prevalence and geographic distribution and to describe clinical features of cryptococcosis patients.

METHODS

We used the 2016–2022 Merative MarketScan Commercial/Medicare and Multi-State Medicaid Databases (https://www.merative.com/documents/brief/marketscan-explainer-general). The Commercial/Medicare database includes health insurance claims data submitted by large employers and health plans for >54 million employees, their dependents, and retirees with employer-provided Medicare Supplemental and Medicare Advantage plans throughout the United States. The Medicaid database includes similar information from >16 million patients across several geographically dispersed states.

We identified patients assigned the International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM), code B45 for cryptococcosis during April 1, 2016–April 1, 2022. The index date was the date this code was first used during the study period. Patients for whom cryptococcosis was listed on a laboratory or radiology claim alone (10% in the commercial database and 3% in Medicaid) were excluded to remove rule-out diagnoses. We selected cryptococcosis patients with continuous insurance enrollment in the 90 days before and after their first B45 code within the study period, estimated prevalence (the denominator was patients with continuous insurance enrollment in the 90 days before and after their first health care visit in the study window), and examined demographic characteristics, underlying conditions, symptoms, clinical form of cryptococcosis, diagnostic testing, outpatient prescriptions, and complications using ICD-10-CM, International Classification of Diseases, 10th Revision, Procedure Coding System (ICD-10-PCS), and Current Procedural Terminology (CPT) codes (Supplementary Table 1). We also examined the geographic distribution of cryptococcosis patients in the commercial insurance database. Geographic data are unavailable in the Medicaid database, and race/ethnicity data are unavailable in the Commercial/Medicare database.

RESULTS

Cryptococcosis prevalence was 3.4 per 100 000 commercially insured patients (n = 1025/30 399 082) and 6.5 per 100 000 Medicaid patients (n = 603/9 273 266) (Table 1). Prevalence was highest among commercially insured patients aged ≥65 years (10.9 per 100 000) and among Medicaid patients aged 45–54 years (32.1 per 100 000). In both cohorts, prevalence was higher among males vs females (commercial: 4.5 vs 2.3; Medicaid: 8.9 vs 4.5). Among Medicaid patients, prevalence was highest among non-Hispanic Black patients (9.2), followed by those of non-Hispanic other race (4.8). In the commercial insurance data set, 54.3% of cases occurred in residents of the South (rate: 4.1 per 100 000) (Supplementary Figure 1), and 19.9% occurred in rural residents (rate: 4.8 per 100 000 vs 2.9 in nonrural residents).

Cryptococcosis patients with Medicaid (vs those with commercial insurance) more often had HIV (59.0% vs 18.2%) and less often had immunosuppressive medication use (16.4% vs 34.5%), diabetes (15.3% vs 28.9%), cancer (13.9% vs 20.2%), immune-mediated inflammatory diseases (6.0% vs 12.0%), and solid organ or stem cell transplantation (3.6% vs 15.5%) (Table 2).

In both data sets, respiratory signs and symptoms (commercial: 57.3%; Medicaid: 58.5%) and systemic symptoms (commercial: 55.0%; Medicaid: 73.1%) were more frequent than central nervous system–related signs and symptoms (commercial: 20.2%; Medicaid: 34.3%); 54.6% of Medicaid patients vs 34.6% of commercially insured patients were hospitalized on the index date. Meningitis was the most common specified form of cryptococcosis (commercial: 34.3%; Medicaid: 58.5%), followed by pulmonary (commercial: 30.9%; Medicaid: 13.4%).

In both data sets, ∼66% of patients received diagnostic testing. Histopathology (commercial: 35.6%; Medicaid: 24.4%) and cryptococcal antigen (CrAg) testing (commercial: 28.5%; Medicaid: 16.7%) were the most common laboratory tests. Lumbar puncture was more frequently performed among patients with Medicaid (47.8%) vs commercial insurance (34.0%). Antifungal susceptibility testing was performed for 6.4% of commercially insured patients and 4.8% of Medicaid patients. Approximately 60% of patients in both data sets received outpatient antifungal treatment, primarily fluconazole. Among patients with meningitis in both databases, ∼8% developed stroke.

DISCUSSION

This analysis of commercial health insurance and Medicaid claims data provides updated prevalence estimates showing that cryptococcosis remains a substantial public health issue. Our results revealed disparities by insurance type, with prevalence nearly twice as high in Medicaid patients compared with those with commercial insurance. The higher cryptococcosis rates among males and Black patients are consistent with other studies [4, 5, 7, 9] and are likely related to the disproportionate burden of HIV infection among people with lower income, men, and Black/African American persons [10]. The higher observed cryptococcosis rates in the Southern United States may partly reflect the large population of people with HIV and high rates of new HIV diagnoses in the South [10]. The higher cryptococcosis rate among commercial health insurance enrollees in rural areas merits further investigation but could be related to C. gattii exposure in areas abundant with mature trees [11].

The total number of cryptococcosis cases nationwide remains unknown, as public health surveillance for cryptococcosis is limited to only 3 states (Louisiana, Oregon, and Washington) [12]. Using the number of estimated cryptococcosis-associated hospitalizations nationally in 2019 (n = 4900) [7] and the percentage of Medicaid cryptococcosis patients hospitalized in our analysis (55%), a crude estimate of >8900 total cases (both hospitalized and nonhospitalized) per year could be expected. Improved and expanded public health surveillance would allow for a better estimate of the number of cases, monitoring trends and emergence of C. gattii into areas where it has not previously been documented, and identifying emerging patient risk groups.

Patients with HIV represented one-third of all cryptococcosis patients in this analysis; in recent single-center studies, this proportion ranged from 20% (Kentucky) to 63% (New York) [3, 13–15]. Transplant recipients are also considered at high risk, representing 5%–20% of cryptococcosis patients [3, 4, 13, 14]; our results (11%) are consistent with previous studies. Our analysis also found a substantial proportion of patients (29%) without any of the immunosuppressive conditions we examined. Diagnostic delays [8, 15], substantial rates of neurological sequelae [16], and higher mortality in non-HIV patients compared with HIV patients [4, 8, 9, 15, 17] indicate the importance of early suspicion for cryptococcosis even among patients without traditional risk factors.

Although the total proportion of patients who received any diagnostic testing for cryptococcosis was similar in commercially insured vs Medicaid patients (66%), the difference in test types by insurance type (ie, more frequent histopathology and CrAg testing among commercially insured patients and more frequent lumbar punctures among Medicaid patients) may reflect cryptococcal meningitis being more common among people with HIV, whereas cryptococcal pneumonia is the more frequent presentation among people without HIV [4]. Treatment recommendations for most forms of cryptococcosis and most patient populations include induction therapy with amphotericin B plus flucytosine, followed by fluconazole [18, 19]. The data sets we used do not include information about inpatient medications given, which likely explains why the proportion of patients who received antifungal treatment (60%) appears lower than expected.

Although the MarketScan databases are broadly representative of patients with commercial health insurance and Medicaid, they are large convenience samples and not representative of the entire US population. The lack of race/ethnicity information in the commercial data set and the lack of geographic information in the Medicaid data set are also limitations, as is the lack of laboratory test result data (eg, CD4 cell count, cerebrospinal fluid profile, Cryptococcus species). Potential misclassification and underdetection of medical conditions, signs and symptoms, and complications are inherent limitations of medical code–based analyses. Lastly, the data sets used do not contain mortality information; therefore, this analysis is biased toward patients who survived at least 90 days after the index date.

Nonetheless, this analysis provides an updated prevalence estimate of cryptococcosis in the United States. Given its potential for severe disease and poor outcomes, our results suggest the importance of considering cryptococcosis as a cause of pulmonary infection and meningitis, particularly in the Southern United States, both in patients with HIV and in those with certain other immunocompromising conditions. More research is needed to better quantify the risk for cryptococcosis among different patient populations. Expanded and strengthened public health surveillance for cryptococcosis could help to fill this gap.

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.

Acknowledgments

The authors thank Malavika Rajeev for assistance with Supplementary Figure 1.

Disclaimer. The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

Patient consent. This study did not include factors necessitating patient consent.

Ethical approval. This activity was reviewed by the CDC and was conducted consistent with applicable federal law and CDC policy (eg, 45 C.F.R. part 46.102(l)(2), 21 C.F.R. part 56; 42 U.S.C. §241(d); 5 U.S.C. §552a; 44 U.S.C. §3501 et seq.).

Financial support. No specific funding was received for this work.

References

Author notes