Serious outcomes of medically attended, laboratory‐confirmed influenza illness among school‐aged children with and without asthma, 2007‐2018

Abstract Background Asthma was associated with influenza hospitalizations in children during the 2009 pandemic, but it is unclear if asthma is associated with serious illness during seasonal epidemics. Little is known regarding the effect of vaccination on influenza severity in children with asthma. Methods Children aged 5‐17 years in a community cohort presenting with acute respiratory illness were prospectively enrolled and tested for influenza from 2007‐08 through 2017‐18 (excluding the 2009‐10 pandemic season). Data from the electronic health record were extracted to determine asthma status and serious outcomes associated with influenza infection. A serious outcome was defined as hospitalization, emergency department visit, and/or pneumonia diagnosis within 30 days of symptom onset. Multivariable logistic regression models were used to assess asthma status and effect of vaccination on odds of a serious outcome. Results One thousand seven hundred and sixty four medically‐attended influenza infections among school‐aged children were included. Asthma was confirmed in 287 (16%) children. A serious influenza‐associated outcome occurred in 104 (6%) children. The odds of a serious outcome did not differ between those with confirmed asthma and those without asthma [adjusted odds ratio (aOR): 1.35, 95% confidence interval (CI): (0.77‐2.35), P = .3]. The effect of vaccination on serious outcomes was not modified by asthma status [aOR for children without asthma: 0.55 (95% CI: 0.28‐1.07), children with asthma: 1.39 (95% CI: 0.53‐3.69); interaction P‐value = .12]. Conclusions Asthma was not a risk factor for serious illness among children with influenza. Additional studies are needed to better understand the role of influenza vaccination in preventing serious outcomes among children with asthma.


| INTRODUC TI ON
Each winter, influenza epidemics cause substantial disease burden for persons of all ages, with an estimated 9-35 million illnesses resulting in up to 708 000 influenza-associated hospitalizations in the United States. 1 School-aged children typically have the highest attack rates, and children with asthma may be at greater risk for developing severe illness and outcomes associated with influenza infection. [2][3][4][5] Among children hospitalized with influenza, asthma was the most common underlying condition reported, particularly during the 2009 H1N1 pandemic. [6][7][8] Some studies have identified asthma as a risk factor for more severe influenza illness, including asthma exacerbations and need for intensive care. This was observed among children hospitalized during the 2009 pandemic, but not consistently among children hospitalized with seasonal influenza infection. [9][10][11][12] Annual influenza vaccination was first recommended by the US Advisory Committee on Immunization Practices for all children aged 6-23 months in 2004; by 2008, annual influenza vaccination was recommended for all children aged 6 months-18 years. 13 However, recommendations for vaccination of persons in certain groups at higher risk for influenza complications, including asthma, have been in place since the 1960s. 14 Despite the long-standing recommendations, data on the effect of influenza vaccination in children with asthma are limited. Most studies use non-specific outcomes or do not include laboratory-confirmed influenza infection. [15][16][17] One study examined vaccine effectiveness during the 2009 pandemic and found no differences in vaccine effectiveness among those with and without asthma. 18 Similar findings were reported for children and adults with and without asthma during seasonal epidemics, although estimates were lower among those with asthma. 19 Asthma may have contributed to influenza severity in children during the 2009 pandemic, but it is unclear whether there is a similar increased risk during seasonal epidemics. It is not known whether influenza vaccination reduces the risk of serious influenza illness among children with asthma. In this analysis, we examined whether school-aged children with asthma were at an increased risk of serious outcomes associated with medically attended, laboratory-confirmed influenza illness across 10 seasonal epidemics and whether the effect of influenza vaccination on serious outcomes differed between children with and without asthma.

| Study population and procedures
This is a secondary analysis of data from annual population-  [20][21][22][23][24][25][26][27][28] In this community, residents receive nearly all outpatient and inpatient care from the Marshfield Clinic Health System, where both inpatient and outpatient records are accessible. 29   , was ≤7 days in duration, and they had not received antiviral medication before enrollment.
This analysis was restricted to children aged 5-17 years with influenza illness confirmed by reverse transcription-polymerase chain reaction (RT-PCR). Those with a co-infection or unsubtypeable influenza A infection were excluded ( Figure S1).
At enrollment, parents/guardians reported on their child's illness onset and symptoms, household exposure to smoke (except in 2007-08), and race/ethnicity. A respiratory sample (nasopharyngeal or combined nasal and oropharyngeal swabs, depending on season) was obtained and tested for influenza by RT-PCR, including subtype. In addition, information on diagnosis codes, outpatient visits, emergency department (ED) visits, hospitalizations, and medications was extracted from the electronic health record to determine high-risk status, 19 number of outpatient visits in the past 12 months (categorized as 0, 1-4, and ≥5), and receipt of a prescription for antivirals in the 7 days after illness onset. These data were also used to determine asthma status and related serious outcomes, described below.

| Influenza vaccination
Influenza vaccination history for the enrollment season and all prior seasons was obtained from the immunization registry used by vaccine providers serving the study population. 30 Children aged ≥9 years were considered vaccinated if they received an influenza vaccine during their enrollment season >14 days before their illness onset. Children aged 5-8 years were considered fully vaccinated if they (a) received two doses at least 28 days apart during their enrollment season and >14 days before their illness onset or (b) one dose during their enrollment season >14 days before illness onset and ≥2 doses in prior seasons. 31 Children aged 5-8 years vaccinated with one dose during their enrollment season >14 days before illness onset and with <2 prior doses were considered partially vaccinated.
Participants vaccinated within 14 days prior to symptom onset were excluded ( Figure S1).

| Asthma
Asthma status was based on asthma-specific diagnosis codes (ICD-9 code 493.* prior to October 1, 2015, and ICD-10 code J45.* on or after October 1, 2015) and prescription medications for asthma in the 2 years prior to enrollment.
Step-up medications and medication classes considered consistent with a diagnosis of mild persistent or more severe asthma were assessed (Table S1).
Step 1 asthma medications included short-acting beta-agonists.
Prescriptions for oral and injectable steroids were not included because they are frequently prescribed for non-asthma diagnoses.
Asthma status was classified into three groups: confirmed, probable, and no asthma. Children with confirmed asthma had an asthma-specific diagnosis code and a prescription for a step 2 or higher asthma medication. Children classified with probable asthma had either (a) an asthma-specific diagnosis code or a prescription for a step 2 or higher asthma medication or (b) an asthma-specific diagnosis code and a prescription for a step 1 asthma medication. Children without an asthma-specific diagnosis code and no prescription for a step 1 or higher asthma medication were classified as no asthma.
Those without an asthma-specific diagnosis code, but had a prescription for a step 1 asthma medication, were excluded ( Figure S1).

| Serious outcome
A serious outcome was defined as hospitalization, ED visit, and/or pneumonia diagnosis within 30 days of symptom onset. A dichotomous composite variable was used for the analysis, and a child with at least one outcome of interest was classified as having a serious outcome. Notes from hospitalizations and ED visits were reviewed by a physician (SCO) to determine whether or not the visits were related to the influenza illness episode. Children were classified as having pneumonia if they received a diagnosis code for pneumonia and a new prescription for antibiotics within 7 days of the pneumonia diagnosis. Those with a code for pneumonia, but no new prescription for antibiotics within 7 days, were excluded ( Figure S1).

| Statistical methods
Only the first confirmed influenza illness per season per child was  Asthma was confirmed in 287 (16%) children, and 227 (13%) had probable asthma ( Figure S1). Children with confirmed asthma differed from children without asthma with regard to several characteristics (Table 1). Children with asthma were more likely to be male (60% vs 49%), have a high-risk condition other than asthma (13% vs 6%), have ≥5 outpatient visits in the previous year (59% vs 35%), and be vaccinated (42% vs 24%). At the time of enrollment, symptoms reported more often by influenza cases with asthma (vs no asthma) included shortness of breath (49% vs 30%) and wheezing (44% vs 24%). Children with influenza and asthma were more likely to receive antiviral treatment compared with those without asthma (22% vs 7%).

| Study population
Children with probable asthma were less likely than children with confirmed asthma to have ≥5 outpatient visits in the previous year (42% vs 59%), be vaccinated (32% vs 42%), receive antivirals (12% vs 22%), and report shortness of breath (37% vs 49%). There were no differences between children with confirmed asthma, probable asthma, or no asthma with regard to race/ethnicity, Medicaid TA B L E 1 Characteristics of school-aged children with influenza by asthma status  Children with a serious outcome were more likely to be 5-8 years old (59% vs 41%, P = .0003) and have ≥5 outpatient visits in the previous year (55% vs 38%) (

| Asthma and effect of vaccination on serious outcomes
The association between vaccination and serious outcomes was not modified by asthma status after adjusting for influenza type/ subtype, age, sex, and presence of a high-risk condition other than asthma. Among children without asthma, the adjusted odds ratio for a serious outcome in vaccinated vs unvaccinated children was 0.55 (95% CI: 0.28-1.07). Among asthmatic children, this odds ratio was 1.39 (95% CI: 0.53-3.69) (interaction P = .12). Similar results were obtained in two sensitivity analyses. The first one excluded partially vaccinated children, and the second one included children who met the criteria for possible asthma (Table S2). pandemic. 32 Asthma status was not associated with increased odds of a serious outcome in children with influenza. This is consistent with studies among children hospitalized with laboratoryconfirmed influenza during seasonal epidemics where asthma was not associated with severe illness. 9,33 Additionally, vaccination did not modify the risk of a serious outcome in children with asthma.

| D ISCUSS I ON
This supports findings from the US Flu VE Network showing vaccine effectiveness estimates did not differ among children with and without asthma. 19 An increased risk for influenza-related complications among children with asthma was found in a primary care cohort with influenza or influenza-like illness during the 2009 pandemic, after adjusting for confounders such as age and vaccination status. 32 Increased asthma exacerbation and serious illness were noted among children hospitalized with pandemic A(H1N1)pdm09 in multiple studies. 9,10,34,35 In the current study of seasonal influenza, the association between asthma and serious outcomes became significant in a sensitivity analysis that included children with probable asthma, suggesting a potential true association that may be detected in larger samples. It is also possible that serious outcomes among children with asthma may be influenza type/ subtype specific. In addition to data from the pandemic, another study found more frequent asthma exacerbations among children with influenza A vs influenza B infection. 34 We did not observe

| CON CLUS IONS
Asthma is the most common underlying condition among children hospitalized with influenza 5,6 and is often cited as a risk factor for hospitalization among both children and adults with influenza. In this multi-season study, asthma was not associated with a serious outcome among children with influenza, and influenza vaccination did not modify the risk of a serious outcome. However, asthma is common among children hospitalized for influenza, and influenza-attributable hospitalizations occur more often among children aged 6-23 months with asthma compared to those without. 5 Additional studies are needed to

TA B L E 2 (Continued)
better understand the role of influenza vaccination in preventing these serious outcomes among children with asthma. In particular, children with severe or poorly controlled asthma merit further investigation to understand the effects of influenza illness and vaccination.

ACK N OWLED G EM ENTS
The authors gratefully acknowledge Jill Ferdinands for her critical review and feedback on the manuscript and Carla Rottscheit for managing the data for this study.

CO N FLI C T O F I NTE R E S T
HQM and KEH report research support from Seqirus for unrelated studies. All other authors report no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the study conclusions are unavailable for public access because informed consent to share said data (beyond the research team) was not obtained from study participants.