Invasive Aspergillosis After Influenza and Other Viral Respiratory Infections Among Intensive Care Unit Patients in a Commercially Insured Population in the United States, 2013–2018

Abstract

Influenza-associated aspergillosis (IAA) is an emerging phenomenon in intensive care unit patients with severe influenza. In a large US health insurance claims database, IAA was uncommon (0.3%) during 2013–2018. The low IAA frequency likely reflects underdiagnosis and differences in medical practices or epidemiologic differences.

Invasive aspergillosis (IA) is an infection caused by the fungus Aspergillus. Persons with weakened immune systems, such as those with history of solid organ or stem cell transplantation, cancer, receipt of high-dose corticosteroids, immunodeficiencies, or with structural lung disease such as chronic obstructive pulmonary disease appear particularly susceptible to IA, which can cause serious morbidity and mortality in these groups [1]. The overall rates of IA are unknown in the United States, and prevalence varies across different institutions [2–4]. Since the 2009 H1N1 pandemic, studies conducted primarily in Europe have identified IA as a postinfluenza complication in previously healthy persons [5, 6]. In the United States, only a few cases of influenza-associated aspergillosis (IAA) have been described in published literature [7–9]. Limited information exists about whether other noninfluenza viral infections may play a role in developing secondary IA, although the pathophysiology of these secondary infections may be similar to IAA. In the absence of comprehensive surveillance systems and clinical studies in the United States, we used a large health insurance claims database to determine the frequency of IA, IAA, and noninfluenza viral respiratory infection-associated aspergillosis (NIVAA).

METHODS

The 2013–2018 IBM MarketScan Commercial Claims and Encounters and Medicare Supplemental Databases include health insurance claims data for approximately 80 million employees, dependents, and retirees throughout the United States. We accessed the data through Treatment Pathways, a web-based tool that includes data from people whose health insurance plans or employers contribute prescription drug data to MarketScan. MarketScan data capture up to 14 International Classification of Diseases (ICD) diagnosis codes.

We identified patients with intensive care unit (ICU) visits for IA, influenza, and noninfluenza viral respiratory infections, using ICD, 9^th and 10^th Revision, Clinical Modification (ICD-9-CM and ICD-10-CM) codes (Supplemental Table). We limited our analysis to patients who were continuously enrolled in MarketScan during the 3 months before and the 3 months after the diagnosis of interest. We examined the proportion of patients who developed IA in the 2 weeks and 3 months after influenza or noninfluenza viral respiratory infection. We examined demographic characteristics, underlying conditions, hospitalization features, and diagnostic procedures associated with IA. MarketScan data do not include detailed information about clinical features, medications administered in the inpatient setting, or specimen types. Statistical testing was not performed for this descriptive analysis.

Patient Consent Statement

This study does not include factors necessitating patient consent.

RESULTS

Among ~64 million patients with any healthcare encounter during 2013–2018, ~7 million patients had any hospitalization. Of those ~7 million patients, 948 (0.01%) had IA codes recorded. Among ~1.5 million ICU patients, 341 (0.02%) had IA without influenza or another viral respiratory infection (Figure 1). Within a 3-month time window, IA occurred in 25 (0.3%) of 7672 ICU patients with influenza and in 45 (0.3%) of 14 673 ICU patients with noninfluenza viral respiratory infections. When we restrict the time window to 2-weeks, IA occurred in 15 (0.2%) of 7672 ICU patients with influenza and 37 (0.3%) of 14 673 ICU patients with noninfluenza viral respiratory infections.

Figure 1.

Patients included the analysis of invasive aspergillosis in a commercially insured population in the United States, 2013–2018.

The median ages for patients who had IAA, NIVAA, and IA only were similar (56–59 years). However, 16% of patients who had NIVAA, 4% of patients who had IAA, and 4% of patients who had IA only were <18 years old (Table 1). Other demographic characteristics were similar across the 3 patient groups.

Table 1.

Demographic Characteristics, Laboratory Tests, Underlying Conditions, and Hospitalization Features Among Intensive Care Unit Patients With Invasive Aspergillosis, Influenza-Associated Aspergillosis, and Noninfluenza Viral Respiratory Infection-Associated Aspergillosis, United States, 2013–2018

	IA Only	Influenza-Associated Aspergillosis	Noninfluenza Viral Respiratory Infection-Associated Aspergillosis
Characteristic	N = 341	N = 25	N = 45
Median age (range)	59 (0–90)	58 (12–84)	56 (7–82)
Age in Years
0–17	14 (4%)	1 (4%)	7 (16%)
18–34	34 (10%)	2 (8%)	3 (7%)
35–44	19 (6%)	3 (12%)	5 (11%)
45–54	58 (17%)	4 (16%)	7 (16%)
55–64	107 (31%)	7 (28%)	15 (33%)
65+	109 (32%)	8 (32%)	8 (18%)
Male sex	168 (49%)	13 (52%)	20 (44%)
Region	a		b
South	106 (31%)	10 (40%)	14 (31%)
Northeast	56 (16%)	4 (16%)	11 (24%)
Midwest	94 (28%)	5 (20%)	10 (22%)
West	80 (23%)	6 (24%)	9 (20%)
Rural location	30 (9%)a	4 (16%)	5 (11%)b
Laboratory Tests
Culture	74 (22%)	5 (20%)	10 (22%)
Pathology	249 (73%)	15 (60%)	36 (80%)
Cytopathology	206 (60%)	15 (60%)	31 (69%)
Galactomannan test	34 (10%)	1 (4%)	5 (10%)
Other relevant testc	61 (18%)	3 (12%)	6 (13%)
Underlying Conditions
Immune-mediated inflammatory diseases	15 (4%)	2 (8%)	6 (13%)
Chronic obstructive pulmonary disease	186 (55%)	15 (60%)	23 (51%)
Diabetes	77 (23%)	6 (24%)	10 (22%)
Hematologic malignancy	52 (15%)	2 (8%)	9 (20%)
HIV/AIDS	5 (1%)	4 (16%)	2 (4%)
Neutropenia	116 (34%)	5 (20%)	18 (40%)
Solid organ or stem cell transplant	58 (17%)	1 (4%)	11 (24%)
Solid organ malignancy	141 (41%)	4 (16%)	20 (44%)
Bronchiectasisd	63 (20%)	2 (13%)	5 (12%)
None of the above conditions	33 (10%)	5 (20%)	2 (4%)
Median length of hospital stay in days (range)	13 (1–226)	18 (2–111)	17 (1–167)
Discharge status: diede	39 (13%)	1 (6%)	7 (17%)

Abbreviations: AIDS, acquired immune deficiency syndrome; HIV, human immunodeficiency virus; IA, invasive aspergillosis.

^aFive (1%) missing.

^bOne (4%) missing.

^cD2 fungal sequencing or susceptibility study.

^dOnly for patients with index dates during 2013–October 1, 2015 (International Classification of Diseases, Ninth Revision code).

^eOnly for patients with index dates during 2013–2015.

Chronic obstructive pulmonary disease was the most common (51%–60%) underlying condition in all 3 patient groups. Approximately one quarter (22%–24%) of patients had diabetes. Twenty percent of the IA only patients and 12%–13% of other groups had bronchiectasis.

Patients with IAA, NIVAA, and IA only had codes for hematologic malignancy (8%, 20%, 15%, respectively), solid or stem cell transplant (4%, 24%, 17%), and solid organ malignancy (16%, 44%, 41%). Sixteen percent in the IAA group, 4% in the NIVAA group, and 1% in the IA only group had human immunodeficiency virus/acquired immune deficiency syndrome. Eight percent of the IAA group, 13% of NIVAA group, and 4% of IA only group had immune-mediated inflammatory diseases. One fifth (20%) of IAA patients, 10% of IA only patients, and 4% of NIVAA patients had no recorded immunocompromising condition.

Half of IA patients had pathology (51%) or cytopathology (49%) codes, and 8% had a galactomannan test. Codes for corticosteroids were present for approximately half (49%) of the NIVAA group, 48% of the IAA group, and 32% of the IA only group. One fifth (20%) of IAA patients, 40% in NIVAA patients, and 34% of patients who had IA only had neutropenia. In-hospital death was recorded in 17% of the NIVAA cohort, 13% in the IA only cohort, and 6% in the IAA cohort.

DISCUSSION

Influenza-associated aspergillosis was uncommonly documented in this large health insurance claims database, orders of magnitude lower than the prevalence in a multicenter study in Belgium and the Netherlands [10], where 1 in 5 patients with severe influenza had invasive pulmonary aspergillosis. However, given that only a few cases have been previously described in the United States [8, 11], the phenomenon may be more common than currently appreciated [12].

It is notable that 1 in 5 patients with IAA had no immunocompromising condition recorded. Invasive aspergillosis typically occurs in patients with immunocompromising conditions but is being described more frequently among traditionally low-risk patients. The precise pathogenesis of how severe influenza may predispose persons to IA is unknown [13]. However, influenza viral pneumonitis or bacterial infections secondary to influenza, such as Staphylococcus aureus or Streptococcus pneumoniae, can lead to respiratory failure, sepsis, and multiorgan failure [14–17]. Similar to bacterial coinfections, severe influenza may involve damage in the respiratory epithelium, allowing Aspergillus to cause serious infection [18]. A recent review showed that 28% of 128 cases of postinfluenza IA were among previously healthy persons [19], and a multicenter study in Belgium and the Netherlands found that approximately half of the patients with invasive pulmonary aspergillosis and severe influenza lacked documented immunocompromised status [3]. Corticosteroid use may also play a role in predisposing patients to IAA [10]. Given the growing concern for coronavirus disease 2019-associated IA documented in persons regardless of their immune status [20–23], more robust prospective epidemiological studies to understand the underlying risk factors for IA among patients in ICUs are warranted.

Medical practices may contribute to the low rate of IAA observed compared with European countries. Galactomannan antigen testing was uncommonly documented in these patients (<10%), whereas it has been used in other countries to screen patients with Aspergillus, especially in intensive care settings [10, 24, 25]. Wider use of galactomannan in serum and BAL might improve diagnosis of IA in critically ill patients, although clinicians should consider the potential for colonization and false positivity [26–28]. Invasive aspergillosis not diagnosed in the ICU is sometimes identified postmortem, and it is one of the most frequently missed diagnosis in the ICU according to a review of autopsy studies [29]. MarketScan data do not include information on autopsies, although few additional cases are likely to have been detected if autopsy data were available because they are rarely preformed in the United States.

Several important limitations are inherent in these administrative data, which are subject to misclassification, undercoding, and lack of laboratory test results and clinical details. We also lack information on IA severity and whether patients died because of these infections or because of something else. Distinguishing between invasive and noninvasive aspergillosis is particularly difficult based on ICD codes alone [30]. Second, corticosteroid use may have been overestimated in this study because we could not distinguish between topical and systemic steroids. Finally, we did not examine antifungal prophylaxis, which could play a role in rates of IAA among immunocompromised patients but perhaps unlikely to affect non-immunocompromised patients.

CONCLUSIONS

Clinicians in the United States may generally be unaware of the risk of IAA in previously healthy persons [12, 31]. Due to lack of awareness, we may have underestimated the occurrences of viral infection-associated IA because IA would likely not be suspected, relevant laboratory tests for IA would likely not be ordered, and clinical diagnosis and treatment for IA would likely be delayed or not made. Given the stark discrepancy in incidence of IAA reported from some European hospitals and the >5% mortality rates observed in our data, prospective studies are needed to assess the true incidence of IAA in US hospitals. In the meantime, these data further support aspergillosis as a possible cause of infection in patients with or without other respiratory infections, including in those with previously normal immune systems.

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.

Notes

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.

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.