Open access Original research

Risk of pneumonia in asthmatic

children using inhaled corticosteroids: a
nested case-­control study in a
birth cohort
Pragya Shrestha,1,2 Chung-­Il Wi,1,2 Hongfang Liu,3 Katherine S King,4 Euijung Ryu,5
Jung Hyun Kwon,1,6 Sunghwan Sohn,3 Miguel Park,7 Young Juhn  ‍ ‍1,2

To cite: Shrestha P, Wi C-­I, ABSTRACT

Liu H, et al. Risk of pneumonia Strengths and limitations of this study
Background  Inhaled corticosteroids (ICSs) are important
in asthmatic children using in asthma management, but there are concerns regarding
inhaled corticosteroids: a ► Our study was a population-­based birth cohort study
associated risk of pneumonia. While studies in asthmatic
nested case-­control study that used longitudinal data and attempted to ad-
adults have shown inconsistent results, this risk in
in a birth cohort. BMJ Open dress major confounders for the study findings such
2022;12:e051926. doi:10.1136/ asthmatic children is unclear.
as asthma control status, severity, individual-­level
bmjopen-2021-051926 Objective  Our aim was to determine the association of
socioeconomic status and vaccination status.
ICS use with pneumonia risk in asthmatic children.
► Prepublication history and ► Our study had an epidemiological advantage as our
Methods  A nested case-­control study was performed
additional supplemental material study setting is a self-­contained healthcare environ-
in the Mayo Clinic Birth Cohort. Asthmatic children (<18
for this paper are available ment which captures all asthma-­related healthcare
years) with a physician diagnosis of asthma were identified
online. To view these files, utilisations.
please visit the journal online from electronic medical records of children born at Mayo
► Our study had inherent limitations for a retrospective
(http://dx.doi.org/10.1136/​ Clinic from 1997 to 2016 and followed until 31 December
study.
bmjopen-2021-051926). 2017. Pneumonia cases defined by Infectious Disease
► It was difficult to fully disentangle the effect of asth-
Society of America were 1:1 matched with controls without
Received 02 April 2021
ma control and severity status from inhaled cortico-
pneumonia by age, sex and asthma index date. Exposure
Accepted 14 February 2022 steroid (ICS) use as a confounder because laboratory
was defined as ICS prescription at least 90 days prior to
measures such as spirometry are not available.
pneumonia. Associations of ICS use, type and dose (low,
► There was lack of confirmation of compliance with
medium and high) with pneumonia risk were analysed
ICS use among the patients as we used prescription
using conditional logistic regression.
data.
Results  Of the 2108 asthmatic children eligible for the
study (70% mild intermittent and 30% persistent asthma),
312 children developed pneumonia during the study Prevention Program (NAEPP) and Global
period. ICS use overall was not associated with risk of
Initiative for Asthma (GINA) guidelines.4 6
pneumonia (adjusted OR: 0.94, 95% CI: 0.62 to 1.41).
Poorly controlled asthma was significantly associated with
Moreover, the recently updated GINA 2019
the risk of pneumonia (OR: 2.03, 95% CI: 1.35 to 3.05; guidelines7 8 also recommend all patients
p<0.001). No ICS type or dose was associated with risk of with asthma to receive either symptom-­driven
pneumonia. or daily ICS to reduce the risk of asthma exac-
Conclusion  ICS use in asthmatic children was not erbation. Therefore, a large proportion of
associated with risk of pneumonia but poorly controlled children with asthma are exposed to ICS.
asthma was. Future asthma studies may need to While ICSs in asthma have shown a good
include pneumonia as a potential outcome of asthma efficacy and safety profile,6 7 there have been
management. debates regarding the associated risk of
pneumonia with long-­term use.9–12 Evidence
© Author(s) (or their supporting risk of pneumonia in patients
employer(s)) 2022. Re-­use
INTRODUCTION with chronic obstructive lung disease appears
permitted under CC BY-­NC. No
commercial re-­use. See rights Asthma is the most common chronic illness to be stronger, however, little is known about
and permissions. Published by affecting 9.6%–13% of children1 2 and one the risk of pneumonia associated with ICS
BMJ. of the five most burdensome diseases among use among asthmatic children and previous
For numbered affiliations see adults in the USA.3 4 Approximately 60% of studies showed inconsistent results.13–24
end of article. the children with asthma in the USA have Systematic reviews and meta-­
analysis
Correspondence to
persistent asthma5 requiring use of inhaled performed on randomised controlled trials
Dr Young Juhn; corticosteroids (ICSs) as the primary control (RCTs) looking at this association have
​juhn.​young@​mayo.​edu therapy per National Asthma Education and extrapolated the total number of pneumonia

Shrestha P, et al. BMJ Open 2022;12:e051926. doi:10.1136/bmjopen-2021-051926 1

Open access

events on the basis of reported adverse events13 14 rather research authorisation, (3) insufficient medical records
than looking at pneumonia as a primary end point. for determining case and exposure status (eg, less than
Most observational studies have used International Clas- two visits other than delivery during the first 2 years of
sification of Diseases (ICD) coding algorithms to identify life), (4) history of chronic diseases making it difficult to
pneumonia diagnosis among asthmatics, which poses a discern asthma status (eg, prematurity, bronchopulmo-
misclassification bias. Similarly, studies from administra- nary dysplasia, immunodeficiency, malignancy, pulmo-
tive databases lack asthma-­related information such as nary fibrosis, bronchiectasis, cystic fibrosis, primary ciliary
asthma severity, asthma control status, ICS dosing (as per dyskinesia, alpha 1 anti-­trypsin deficiency diagnosed by
age group) and other confounders such as vaccination ICD9/ICD10 codes).
status and socioeconomic status (SES) of the patient. Asthma status (both intermittent and persistent
Better knowledge regarding the safety of ICS, specif- asthma) was initially screened using ICD9/ICD10 (493/
ically on the association of ICS use with the risk of J45) and confirmed by manual chart review for a physi-
pneumonia in asthmatic children and potential factors cian diagnosis of asthma. Only subjects with confirmed
accounting for such association such as asthma severity or asthma diagnosis by a physician were considered for this
control status, will help clinicians and parents’ adherence study.
with asthma guidelines as it can reassure clinicians and
parents regarding safety concerns. Case ascertainment (pneumonia)
The aim of our study was to examine the association of All eligible asthmatic children in the birth cohort were
ICS use in paediatric patients with asthma of the Mayo initially screened for pneumonia diagnoses 90 days
Clinic Birth Cohort. Knowledge gained from this study after their physician diagnosis date of asthma through
will help to provide an important insight into the nature 31 December 2017, using validated ICD9/ICD10 codes
of impact of ICS on the risk of pneumonia and mitigate (480–487.0/A37, J09.X1 and J10–J18)35 and confirmed
the potential impact of parental safety concerns about through manual chart review (using Infectious Disease
ICS use on adherence to asthma management recom- Society of America guidelines).36 37 There was a 25%
mending ICS use.25 false positivity of pneumonia by ICD codes alone. The
diagnosis of community-­ acquired pneumonia requires
a patient with a clinically compatible syndrome (fever
METHODS and cough with or without dyspnoea/sputum produc-
Study design and setting tion).36 According to the guidelines, as pneumonia is
This was a nested case-­control study from a subset (the primarily regarded as a clinical diagnosis (without defin-
Mayo Clinic Birth Cohort) of the Olmsted County Birth itive requirement of chest X-­ray for diagnosis and treat-
Cohort. Olmsted County, southeastern Minnesota, is a ment),37 chest X-­ray was not required to be pneumonia
virtually self-­
contained healthcare environment (only cases for this study38 39 according to guidelines. However,
two healthcare systems provide clinical care to nearly we performed sensitivity analysis among the subset of
all Olmsted County, Minnesota residents), and 98% of pneumonia cases with positive chest X-­ ray repeating
residents authorise their medical records to be used for unadjusted and adjusted analyses among that subset.
research.26 According to US census data in 2010, the age, For children with multiple episodes of pneumonia, we
sex and ethnic characteristics of Olmsted County resi- included only the first episode.
dents were similar to those of the state of Minnesota and
the Upper Midwest.27 28 However, Olmsted County has Selection of controls
been becoming more diverse as indicated by the racial Controls were selected from the asthmatic subjects who
and ethnic characteristics of children enrolled in public did not develop pneumonia based on the screening
schools (in 2019, 35.2% reported to be non-­ white).29 described above. Controls were matched 1:1 with cases in
Prevalence of asthma in a population of school-­age chil- term of sex, age and asthma index date (±1 year). Index
dren Olmsted County, Minnesota, in 2000 (17.6%) was date for the control was defined by the date of the clinic
relatively higher than that of children at a national level visit closest to pneumonia event date of the matched case
(12.4%).30 31 (±1 year). Manual chart review was performed to ensure
matched controls did not have diagnosis of pneumonia
Study subjects in the electronic health record (EHR) during the study
Patients <18 years of age were identified through an period.
ongoing National Institute of Health (NIH) R01-­
supported study (HL126667) from a subset of the Exposure status (ICS use)
Olmsted County Birth Cohort (1997–2016) who were Exposure was defined as at least 90 days of ICS use prior
born at Mayo Clinic, Rochester, Minnesota, and received to index date (eg, a diagnosis date of pneumonia for
their primary care at Mayo Clinic throughout the study cases) to have enough duration of exposure to ICS use in
period (1997–2017). The details of this cohort have been relation to the timing of the development of pneumonia.
published previously.32–34 The exclusion criteria included If the ICS was prescribed at least more than 90 days prior
(1) non-­ Mayo birth cohort, (2) individuals without to pneumonia, we assumed patient had been on ICS until

2 Shrestha P, et al. BMJ Open 2022;12:e051926. doi:10.1136/bmjopen-2021-051926

 Open access

pneumonia was developed. This was ascertained using to disseminate the study results by making the published
prescription data from EHRs, not claim data. We chose report available to public freely.
this timeframe for the exposure to ICS by adopting it
from other studies, which assessed an association between Statistical analysis
ICS use and the risk of pneumonia, for fair comparison Descriptive statistics were used to summarise the charac-
(online supplemental figure 1).13 teristics of cases and controls and comparisons were made
All prescriptions for ICS, alone or in combination with with Analysis of Variance (ANOVA) for continuous vari-
other inhalers, dispensed between the asthma index date ables, χ2 for categorical variables and Cochran-­Armitage
and pneumonia diagnosis were identified and classified trend test for ordinal variables. Matched analysis via
into subgroups based on their formulation. Five groups conditional logistic regression accounting for matched
of ICS were identified—(1) beclomethasone metered pairs was used to determine the association between ICS
dose inhaler (MDI), (2) budesonide (dry powder inhaler status as the primary explanatory variable for the risk of
(DPI) or nebules), (3) ciclesonide MDI, (4) fluticasone pneumonia while controlling for potential confounders
(MDI or DPI) and (5) mometasone DPI. Similarly, ICS identified by univariate analysis. A sensitivity analysis was
were grouped into low, medium and high dose depending performed among the subset of pneumonia cases with
on total mcg/mg use each day as defined by the NAEPP positive chest X-­ray repeating unadjusted and adjusted
guidelines (paediatric age-­based ICS dose).6 analyses among that subset. Dosage and type of ICS were
additionally reported between the cases and the controls.
Covariates of interest The literature suggests that the proportion of any ICS use
Other relevant variables were collected from medical among cases (pneumonia) is 0.5 versus 0.35 in controls.13
record review including: demographic variables (age, The minimum number of cases (1:1 matching) to obtain
gender, race/ethnicity), an individual level HOUsing-­ 80% of power48 to detect the difference is 185 subjects.
based SES measure (HOUSES),40 pneumococcal vaccina- Our study has adequate power to address our primary
tion up-­to-­date status41 and influenza vaccination status at study aim for the association between ICS use and the risk
same year as or prior year to the index date42 (based on of pneumonia. All analyses were performed using R statis-
Centers for Disease Control and Prevention guidelines). tical software (V.3.6.2; R Core Team, Vienna, Austria).
HOUSES index is a single factor made up of four items
(number of bedrooms, number of bathrooms, square
footage of the unit and estimated building value of the RESULTS
unit) ascertained from the county assessor’s office by Study subjects
matching subjects’ addresses which were retrieved from The characteristics of the subjects are summarised in
the EHRs with publically available real property data.40 43 table 1. We identified 2108 eligible patients with asthma
HOUSES index has been validated in numerous studies from the birth cohort (n=21 813) of whom 312 children
including asthma and pneumonia-­related outcomes.44 45 with history of pneumonia during the follow-­up period
Asthma exacerbations (defined as oral corticosteroids were identified. Forty-­one per cent were females and
use, emergency department (ED) visit or hospitalisation 26% were non-­White with the mean age of 2.9 years at
for asthma6 46) within 1 year prior to index date were used asthma diagnosis and 12.7 years at the last follow-­up date.
as a surrogate marker for asthma control status, as most of Seventy per cent had intermittent asthma and 30% had
study subjects did not have information for asthma control persistent asthma (23%, 4% and 3% were mild, moderate
status (eg, Asthma Control Test) at the time of pneumonia and severe persistent asthma, respectively).
(no exacerbations vs poorly controlled asthma defined
as the presence of at least one asthma exacerbation).47 ICS use and the risk of pneumonia
The number and frequency of clinic visits was ascer- The results on the associations of ICS use, type and dose
tained as a proxy measure of healthcare access to mini- with pneumonia are summarised in table 2. ICS use
mise a detection bias (differential identification of mild among the cases was 27% whereas ICS use among the
pneumonia by differential healthcare access). Severity of matched controls was 22%. Fluticasone was the most
asthma (intermittent vs persistent) was assessed based on commonly used ICS (20.4%) followed by budesonide
ICS use along with use of other controller medications (2.2 %) among our cohort. Seven subjects were on high
(long-­acting beta adrenergic, leukotriene receptor antag- dose (1.1%), 19 on medium dose (3%) and 129 on low
onist, theophylline and so on) according to the NAEPP dose (20.7%) ICS. ICS use was not associated with risk
guidelines.6 of pneumonia in univariate analysis (OR: 1.30, 95%
CI: 0.89–1.88; p=0.16) (table 3 for univariate analysis).
Patient and public involvement After adjusting for pertinent covariates or confounders
This was a retrospective, nested case-­control study which including number of clinic visits per year, influenza
did not require patient contact, and thus patients were vaccine status and asthma control status, the effect size
not involved in the design of the study nor subject recruit- of ICS use on the risk of pneumonia significantly attenu-
ment. We will follow the NIH research publication policy ated in a way resulting in a different direction, although

Shrestha P, et al. BMJ Open 2022;12:e051926. doi:10.1136/bmjopen-2021-051926 3

Open access

Table 1  Demographic data for pneumonia case and control subjects

Variables Cases (n=312) Controls (n=312) Total (n=624) P value
Female, n (%) 128 (41) 128 (41) 256 (41) 1.000*
Race, n (%)
 White 230 (73.7) 231 (74.0) 461 (73.9) 0.153*
 Black 35 (11.2) 36 (11.5) 71 (11.4)
 Others 46 (14.7) 38 (12.2) 84 (13.4)
 Unknown 1 (0.3) 7 (2.2) 8 (1.3)
Age at asthma diagnosis (years), mean (SD) 2.9 (2.4) 2.9 (2.3) 2.9 (2.3) 0.976†
Average number of clinic visits per year, mean (SD) 9.5 (7.3) 7.6 (4.7) 8.5 (6.2) <0.001†
Influenza vaccine up-­to-­date status, n (%) 219 (70.2) 202 (64.7) 421 (67.5) 0.146*
Pneumonia vaccine up-­to-­date status, n (%) 222 (71.2) 221 (70.8) 443 (71.0) 0.930*
HOUSES‡, n (%) 0.828§
 Q1 (the lowest) 100 (32.7) 98 (32.2) 198 (32.5)
 Q2 69 (22.5) 78 (25.7) 147 (24.1)
 Q3 63 (20.6) 60 (19.7) 123 (20.2)
 Q4 (the highest) 74 (24.2) 68 (22.4) 142 (23.3)
Asthma severity, n (%) 0.226§
 Intermittent 217 (69.6) 234 (75.0) 451 (72.3)
 Mild persistent 72 (23.1) 57 (18.3) 129 (20.7)
 Moderate persistent 13 (4.2) 16 (5.1) 29 (4.6)
 Severe persistent 10 (3.2) 5 (1.6) 15 (2.4)
Asthma status, n (%) <0.001*
 Well-­controlled 192 (61.5) 238 (76.3) 430 (68.9)
 Poorly controlled 120 (38.5) 74 (23.7) 194 (31.1)

*Pearson’s χ2 test.
†Linear model ANOVA.
‡HOUSES: an individual-­level socioeconomic status measures based on real property data.
§Cochran Armitage trend test.
ANOVA, Analysis of Variance; HOUSES, HOUsing-­based socioeconomic status.

statistically not significant (OR: 0.94, 95% CI: 0.62 to 1.41; by positive chest X-­ray findings (133/312 (43%)) which
p=0.75) (table 4 for multivariate analysis). did not affect the overall results and interpretation for
the association (adjusted OR: 0.87, 95% CI: 0.47 to 1.61;
Other risk factors for pneumonia among children with asthma p=0.65) (online supplemental table 1).
Table 2 shows different types or dose of ICS were not
associated with the risk of pneumonia (p=0.63 and 0.43,
respectively). Both cases and controls were adequately
vaccinated with influenza vaccine (70% for cases and 65% DISCUSSION
for controls) and pneumococcal vaccine (71% each for Our study results showed ICS use was not associated with
cases and controls), thus vaccination did not affect risk the risk of pneumonia in asthmatic children in our birth
of pneumonia (p=0.15 and p=0.93, respectively). While cohort. However, poorly controlled asthma (ie, defined
association of asthma severity with pneumonia was not by ICS use, asthma-­related ED visits or hospitalisations)
found to be statistically significant, poorly controlled was significantly associated with the risk of pneumonia.
asthma status within the past year posed the highest risk Our study finding on the association of poorly
of pneumonia, which remained significant after adjusting controlled asthma with the risk of pneumonia is note-
for other covariates (OR: 2.03, 95% CI: 1.35 to 3.05; worthy. As a potential mechanism, poorly controlled
p<0.001). asthma might cause impairment of innate immunity
Given the concern about a potential misclassification function and epithelial barrier disruption and which lead
bias stemming from clinical definition of pneumonia to susceptibility to infection such as pneumonia.49 For
with chest X-­ray finding, we performed sensitivity analysis example, while impairment of rhinovirus-­induced type I
among a subgroup of cases with confirmed pneumonia and III interferon secretion and its subsequent increased

4 Shrestha P, et al. BMJ Open 2022;12:e051926. doi:10.1136/bmjopen-2021-051926

 Open access

Table 2  ICS use for pneumonia cases and control subjects

Medication use Cases (n=312) Controls (n=312) Total (n=624) P value
ICS use, n (%) 85 (27.2) 70 (22.4) 155 (24.8) 0.165*
Type of ICS, n (%) 0.627*
 None 227 (72.8) 242 (77.6) 469 (75.2)
 Beclomethasone 5 (1.6) 3 (1.0) 8 (1.3)
 Budesonide 9 (2.9) 5 (1.6) 14 (2.2)
 Fluticasone 67 (21.5) 60 (19.2) 127 (20.4)
 Mometasone 4 (1.3) 2 (0.6) 6 (1.0)
Dose of ICS, n (%) 0.432†
 None 227 (72.8) 242 (77.6) 469 (75.2)
 Low 70 (22.4) 59 (18.9) 129 (20.7)
 Medium 10 (3.2) 9 (2.9) 19 (3.0)
 High 5 (1.6) 2 (0.6) 7 (1.1)
Non-­ICS controller (biologics, cromolyn, LTRA), n (%) 25 (8.0) 21 (6.7) 46 (7.4) 0.540*

*Pearson’s χ2 test.
†Cochran Armitage trend test.
ICS, inhaled corticosteroid; LTRA, leukotriene receptor antagonists.

replication of rhinovirus have been widely recognised,50 or ICS and formoterol combination as a safer option
such phenomena were not observed in well controlled for better long-­term prognosis.8 Among the cases of our
asthma51 but those with severe therapy resistant atopic study cohort (312 with pneumonia), 217 (70%) were
asthma.52 This might be potentially applicable to bacte- actually intermittent asthma who had not used ICS or
rial infection. For example, Habibzay et al reported that any other controller. With this updated asthma guideline
using the intranasal house-­ dust mite-­ sensitised mouse (ie, intermittent use of ICS or ICS combination), risk of
model of allergic airway disease, an inflammatory pneumonia among intermittent asthma may be reduced
response impaired innate immune function (a reduc- through optimal asthma control. More recently, there
tion in neutrophil recruitment to the airspaces) and led is an increased understanding that patients considered
to bacterial invasion and dissemination.53 Tight junction to have mild asthma have greater morbidity than previ-
formation and transepithelial electrical resistance were ously appreciated.59 This has further necessitated the
significantly lower in epithelial cultures from asthmatic guidelines to recommend ICSs as a controller treatment
donors than from normal controls suggesting that the option for all patients.7 This postulation has been further
bronchial epithelial barrier in asthma is compromised.54 supported in our study by the significant reduction of the
Given our study results on the associations of poorly effect size of the association between ICS use and the risk
controlled asthma status with the risk of pneumonia and of pneumonia after controlling for asthma control status.
the reported under-­treatment of asthma with ICS,55 56 this Despite a relatively small sample size in our study, given
association does raise an important concern of asthmatic the sufficient statistical power to detect the reported
children possibly being suboptimally managed for asthma, effect size for the association between ICS use and the risk
resulting in a higher risk of pneumonia. Also, previously of pneumonia, our study results are unlikely to be subject
claimed ICS use on the risk of pneumonia might have to type II error. Regardless of statistical power or sample
been largely stemming from the inadequately controlled size, if ICS use had been associated with the risk of pneu-
asthma as a major confounder in the literature. Appre- monia, one would expect higher ICS dose to have higher
hension towards ‘stepping-­ up’ the asthma treatment risk of pneumonia, which was not observed in our study.
with ICS may deter adequate control of asthma and lead Future asthma studies should consider including pneu-
to a worse outcome for patients with poorly controlled monia as a potential outcome of asthma control status as
asthma. While a short-­acting beta2-­agonist (SABA) is a the current outcomes of asthma studies largely focus on
historic asthma medication for symptom relief, extensive asthma control, risk of exacerbation and lung function.46
literature and data suggest that its sole use of SABA is asso- Our study results are consistent with the findings
ciated with poor long-­term asthma control with a higher reported by Cazeiro et al based on a meta-­analysis for
risk for asthma exacerbation43 57and mortality.58 More- children with asthma.19 The meta-­analysis of nine trials
over, recently updated GINA guidelines have suggested that revealed at least one event of pneumonia showed a
removing SABA monotherapy as a first line of treatment reduced risk of pneumonia in patients taking ICS (risk
approach among patients with low symptom burden and ratio (RR): 0.65, 95% CI: 0.44 to 0.94). However, the
replacing this with controller medications such as ICS meta-­analysis including all 31 trials revealed no significant

Shrestha P, et al. BMJ Open 2022;12:e051926. doi:10.1136/bmjopen-2021-051926 5

Open access

multiple randomised controlled studies in a subgroup

Table 3  Univariate analysis for the association between
factors and risk of pneumonia analysis, they found no significant difference in the inci-
dence of pneumonia between fluticasone-­treated versus
Variables OR 95% CI P value
budesonide-­treated groups. Moreover, a protective effect
ICS use of ICS on the risk of pneumonia (not requiring hospital-
 No 1 reference isation) was observed in the study (occurrence of pneu-
 Yes 1.30 0.89 to 1.88 0.163 monia was 0.5% for budesonide and 1.2% for placebo).16
HOUSES* Similarly, a recent clinical trial which tested high-­dose ICS
(quadrupling dose: >1000 µg/day vs low ≤1000 µg/day of
 Q1 (the lowest) 1 reference
beclomethasone) showed no association of high-­dose ICS
 Q2 0.88 0.56 to 1.37 0.569 use with the risk of pneumonia.15 An earlier RCT showed
 Q3 1.02 0.66 to 1.57 0.938 the safety of ICS on the risk of infection by comparing ICS
 Q4 (the highest) 1.09 0.71 to 1.69 0.690 treatment versus placebo. The results, indeed, showed a
Average number of 1.09 1.04 to 1.14 <0.001 significant decrease from before to after therapy in the
clinic visits per year percentage of days of upper respiratory tract infection
Influenza vaccine up-­ (21% to 10% ICS vs 19% to 16% placebo) and lower
to-­date status respiratory tract infection (LRTI) (30% to 15% ICS vs
27% to 21% placebo).60
 No 1 reference
On the other hand, some studies have shown the oppo-
 Yes 1.29 0.92 to 1.82 0.142
site effect.13 14 61 62 In a nested case-­control study found an
Pneumonia vaccine increased risk of pneumonia and LRTI with only flutica-
up-­to-­date status sone use (OR: 1.58, 95% CI: 1.29 to 1.93) in adult patients
 No 1 reference with asthma.13 The study used administrative codes (ICD9
 Yes 1.05 0.57 to 1.94 0.876 codes) for case ascertainment, however, did not account
Asthma severity for asthma control status, vaccination status and frequency
of clinic visits. A retrospective study by Qian et al using
 Intermittent 1 reference
Quebec health insurance database found an increased
 Mild persistent 1.39 0.92 to 2.09 0.121
risk of pneumonia associated with current ICS use (RR
 Moderate persistent 0.87 0.39 to 1.94 0.737 1.83, 95% CI: 1.57 to 2.14) with incremental dose relation
 Severe persistent 2.01 0.68 to 5.92 0.205 (RR 1.60 for low dose, RR 1.53 for medium dose and RR
Asthma status 2.67 for high dose ICSs).14 The study again did not clearly
 Well-­controlled 1 reference define and adjust for major confounders such as asthma
control or severity status and others such as vaccination
 Poorly controlled 2.28 1.54 to 3.37 <0.001
history and SES. Moreover, the study only took account
*HOUSES: an individual-­level socioeconomic status measures of hospitalised cases of pneumonia. In summary of the
based on real property data. literature, prospective studies based on RCT or meta-­
ICS, inhaled corticosteroid.
analysis on the association between ICS use and the risk of
pneumonia do not support such association while retro-
difference in the risk of pneumonia between the ICS and spective studies based on administrative data claim such
placebo groups (risk difference: −0.1%; 95% CI: −0.3% association raising study design issues as described above
to 0.2%). Also, O’Byrne et al reported no significantly given the inherent limitations of retrospective studies
increased risk of pneumonia with ICS use (HR=1.29, based on administrative claims data. These retrospective
95% CI: 0.53 to 3.12) based on a 3-­month double-­blind, studies could get an analytic benefit from application of a
placebo-­controlled trial.16 Although pneumonia was validated asthma control index using claim data in paedi-
not the primary outcome in the study (reported as an atric population which might potentially account for the
adverse outcome) and results were based on pooling association between ICS use and the risk of pneumonia.63

Table 4  Association between use of ICS and risk of pneumonia from a multivariate conditional logistic regression model*
Variables Adj. OR 95% CI P value
ICS use, yes versus no 0.94 0.62 to 1.41 0.75
Average number of clinic visits per year 1.07 1.02 to 1.12 0.003
Influenza vaccine up-­to-­date status, yes versus no 1.08 0.75 to 1.56 0.68
Asthma status, poorly controlled versus well-­controlled 2.03 1.35 to 3.05 <0.001
*Model included, any ICS Use, average number of clinic visits per year, influenza vaccine up-­to-­date status and asthma control status.
Adj. OR, adjusted OR; ICS, inhaled corticosteroid.

6 Shrestha P, et al. BMJ Open 2022;12:e051926. doi:10.1136/bmjopen-2021-051926

 Open access

Our study is a community-­based birth cohort study that Author affiliations

1
used rigorous approaches for ascertaining ICS use and Precision Population Science Lab, Department of Pediatrics and Adolescent
Medicine, Mayo Clinic, Rochester, Minnesota, USA
pneumonia incidence through longitudinal data with 2
Pediatric and Adolescent Medicine, Mayo Clinic, Rochester, Minnesota, USA
definite follow-­up period of paediatric patients at Mayo 3
Artificial Intelligence and Informatics, Mayo Clinic, Rochester, Minnesota, USA
Clinic, Rochester, Minnesota, USA. This is a first paedi- 4
Clinical Trials and Biostatistics, Mayo Clinic, Rochester, Minnesota, USA
5
atric population-­based birth cohort study demonstrating Computational Biology, Mayo Clinic, Rochester, Minnesota, USA
6
ICS use not to be associated with risk of pneumonia in Pediatrics, Korea University Medical Center, Seoul, Republic of Korea
7
Division of Allergic Diseases, Mayo Clinic, Rochester, Minnesota, USA
asthmatic children. The study has also attempted to
control for major confounders like asthma control status, Contributors  PS conceptualised and designed the study, designed the data
severity, individual-­level SES and vaccination status (pneu- collection instrument, collected and interpreted the data, drafted the initial
monia and influenza) which are inadequately addressed manuscript, and reviewed and revised the manuscript. C-­IW designed the study,
by previous retrospective studies. collected and interpreted the data, reviewed and revised the manuscript. KSK
designed the data collection instruments, carried out the analyses, and reviewed
Our study has inherent limitations as a retrospective and revised the manuscript. ER conceptualised and designed the study, designed
study. While our case ascertainment for pneumonia was the data collection instrument, supervised data analyses, and reviewed and revised
consistent with current guidelines,38 64 we did not include the manuscript. JHK designed the data collection instrument, reviewed and revised
chest X-­ray finding in pneumonia case ascertainment the manuscript. SS, HL, MP conceptualised and designed the study, reviewed and
revised the manuscript. YJ (guarantor) conceptualised and designed the study,
which might result in a misclassification bias. To address supervised data collection and analysis, interpreted the data, and reviewed and
this concern, we performed sensitivity analysis among a finalised the manuscript. All authors approved the final manuscript as submitted
subgroup of cases with confirmed pneumonia by posi- and agree to be accountable for all aspects of the work.
tive chest X-­ray findings (133/312 (43%)) which did not Funding  All phases of this study were supported by National Institutes of Health
affect the overall results and interpretation for the asso- (NIH)-­funded R01 grant (R01 HL126667), R21 grant (R21AI116839) and R21 grant
ciation (adjusted OR: 0.87, 95% CI: 0.47 to 1.61; p=0.65) (R21 AG65639), Dr Shrestha is funded under NIH training grant T32 GM008685-­22.
(see online supplemental table 1). As discussed above, Disclaimer  YJ is principal investigator (PI) of the respiratory syncytial virus
incidence study supported by GlaxoSmithKline. The remaining authors have no
it is difficult to fully disentangle the effect of asthma
financial relationships relevant to this article to disclose.
control and severity status from ICS use as a confounder
Competing interests  None declared.
as retrospective studies did not have precise and accurate
measure of asthma control and severity status around Patient consent for publication  Consent obtained from parent(s)/guardian(s).
the time of pneumonia (eg, Asthma Control Test score, Ethics approval  We obtained the approval from the Mayo Clinic Institutional
lung function measures and so on). Also, we were not Review Board (IRB) for the study. The IRB number for approval is 14-­009934.
able to quantify and take into account the exposure to Provenance and peer review  Not commissioned; externally peer reviewed.
ICS or systemic steroid (eg, total ICS or systemic steroid Data availability statement  Data are available upon reasonable request. The
dose during the follow-­up period) as pharmacy claim datasets generated and/or analysed during the current study are not publicly
available as they include protected health information. Access to data could be
data were not available to our study. Also, the proportion discussed per the institutional policy after IRBs at Mayo Clinic approves it.
of patients not on ICS was higher in our cohort, which
Supplemental material  This content has been supplied by the author(s). It has
may have caused decreased exposure to ICS. Another not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been
limitation in our study lies in the lack of confirmation of peer-­reviewed. Any opinions or recommendations discussed are solely those
compliance with ICS use among the patients as we used of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and
prescription data. While this is a limitation of our study responsibility arising from any reliance placed on the content. Where the content
includes any translated material, BMJ does not warrant the accuracy and reliability
as a retrospective study, accurate measurement of compli- of the translations (including but not limited to local regulations, clinical guidelines,
ance in even prospective studies is often challenging as terminology, drug names and drug dosages), and is not responsible for any error
claim or self-­report may not necessarily be accurate. As and/or omissions arising from translation and adaptation or otherwise.
with a retrospective study using medical index search Open access  This is an open access article distributed in accordance with the
codes (eg, ICD codes), the major limitation remains Creative Commons Attribution Non Commercial (CC BY-­NC 4.0) license, which
permits others to distribute, remix, adapt, build upon this work non-­commercially,
on the accuracy of the ICD codes used for asthma and and license their derivative works on different terms, provided the original work is
pneumonia diagnosis. To address this concern, we did properly cited, appropriate credit is given, any changes made indicated, and the use
perform a manual chart review to verify the accuracy of is non-­commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
the coded cases of pneumonia and excluded incorrectly
ORCID iD
coded diagnoses. It is noteworthy that we found misclas- Young Juhn http://orcid.org/0000-0003-2112-4240
sification of pneumonia (false positive rate of 25%) by
manual chart review which may pose an important impli-
cation on interpretation of retrospective studies based
on claim data. REFERENCES
1 Valet RS, Gebretsadik T, Carroll KN. High asthma prevalence and
In conclusion, ICS use in asthmatic children was not increased morbidity among rural children in a Medicaid cohort. Ann
associated with risk of pneumonia but poorly controlled Allerg Asthma Im 2011;106:467–73.
2 Centers for Disease Control and Prevention (CDC). Asthma mortality
asthma was. Future asthma studies may need to include and hospitalization among children and young adults-­United States,
pneumonia as a potential outcome of asthma control 1980-­1993. MMWR Morb Mortal Wkly Rep 1996;45:350–3.
3 Druss BG, Marcus SC, Olfson M, et al. Comparing the national
status. Our findings will need to be verified in larger, economic burden of five chronic conditions. Health Aff
prospective cohort studies. 2001;20:233–41.

Shrestha P, et al. BMJ Open 2022;12:e051926. doi:10.1136/bmjopen-2021-051926 7

 Open access

4 Centers for Disease Control and Prevention (CDC). Vital signs: 31 Yawn BP, Wollan P, Kurland M, et al. A longitudinal study
asthma prevalence, disease characteristics, and self-­management of the prevalence of asthma in a community population of
education: United States, 2001-­2009. MMWR Morb Mortal Wkly Rep school-­age children. J Pediatr 2002;140:576–81. doi:10.1067/
2011;60:547–52. mpd.2002.123764
5 (CDC) CfDCaP. AsthmaStats - asthma severity among children with 32 Seol HY, Rolfes MC, Chung W, et al. Expert artificial intelligence-­
current asthma 2019. based natural language processing characterises childhood asthma.
6 Wea B. Expert panel report 3 (EPR-­3): guidelines for the diagnosis BMJ Open Respir Res 2020;7.
and management of Asthma-­Summary report 2007. J Allergy Clin 33 Wi C-­I, Sohn S, Rolfes MC, et al. Application of a natural language
Immunol 2007;120:S126–7. processing algorithm to asthma ascertainment. an automated chart
7 GINA GIfA. GINA main report, 2019. Available: https://ginasthma.org/​ review. Am J Respir Crit Care Med 2017;196:430–7.
gina-reports/ [Accessed 27 Feb 2020]. 34 Sohn S, Wi C-­I, Wu ST, et al. Ascertainment of asthma prognosis
8 Reddel HK, FitzGerald JM, Bateman ED, et al. GINA 2019: using natural language processing from electronic medical records. J
a fundamental change in asthma management: treatment of Allergy Clin Immunol 2018;141:2292–4.
asthma with short-­acting bronchodilators alone is no longer 35 Drahos J, Vanwormer JJ, Greenlee RT, et al. Accuracy of ICD-­9-­CM
recommended for adults and adolescents. Eur Respir J 2019;53 codes in identifying infections of pneumonia and herpes simplex
doi:10.1183/13993003.01046-2019 virus in administrative data. Ann Epidemiol 2013;23:291–3.
9 Pandya D, Puttanna A, Balagopal V. Systemic effects of inhaled 36 Metlay JP, Waterer GW, Long AC, et al. Diagnosis and treatment
corticosteroids: an overview. Open Respir Med J 2014;8:59–65. of adults with community-­acquired pneumonia. An official
10 Ernst P, Suissa S. Systemic effects of inhaled corticosteroids. Curr clinical practice guideline of the American thoracic Society and
Opin Pulm Med 2012;18:85-­9. infectious diseases Society of America. Am J Respir Crit Care Med
11 Lipworth BJ. Systemic adverse effects of inhaled corticosteroid 2019;200:e45–67.
therapy: a systematic review and meta-­analysis. Arch Intern Med 37 Mandell LA, Wunderink RG, Anzueto A, et al. Infectious Diseases
1999;159:941–55. Society of America/American Thoracic Society consensus guidelines
12 Suissa S, Kezouh A, Ernst P. Inhaled corticosteroids and the risks of on the management of community-­acquired pneumonia in adults.
diabetes onset and progression. Am J Med 2010;123:1001–6. Clin Infect Dis 2007;44:S27–72.
13 McKeever T, Harrison TW, Hubbard R, et al. Inhaled corticosteroids 38 Bradley JS, Byington CL, Shah SS, et al. The management of
and the risk of pneumonia in people with asthma: a case-­control community-­acquired pneumonia in infants and children older than 3
study. Chest 2013;144:1788–94. months of age: clinical practice guidelines by the Pediatric Infectious
14 Qian CJ, Coulombe J, Suissa S, et al. Pneumonia risk in asthma Diseases Society and the Infectious Diseases Society of America.
patients using inhaled corticosteroids: a quasi-­cohort study. Br J Clin Clin Infect Dis 2011;53:e25–76.
Pharmacol 2017;83:2077–86. 39 Harris M, Clark J, Coote N, et al. British Thoracic Society guidelines
15 McKeever T, Mortimer K, Wilson A, et al. Quadrupling inhaled for the management of community acquired pneumonia in children:
glucocorticoid dose to abort asthma exacerbations. N Engl J Med update 2011. Thorax 2011;66:ii1–23.
2018;378:902–10. 40 Juhn YJ, Beebe TJ, Finnie DM, et al. Development and initial testing
of a new socioeconomic status measure based on housing data. J
16 O'Byrne PM, Pedersen S, Carlsson L-­G, et al. Risks of pneumonia in
Urban Health 2011;88:933–44.
patients with asthma taking inhaled corticosteroids. Am J Respir Crit
41 (CDC) CfDCaP. Pneumococcal vaccine recommendations. vaccines
Care Med 2011;183:589–95.
and immunizations web site. Available: https://www.cdc.gov/​
17 Bansal V, Mangi MA, Johnson MM, et al. Inhaled corticosteroids and
vaccines/vpd/pneumo/hcp/recommendations.html [Accessed 2 Aug
incident pneumonia in patients with asthma: systematic review and
2020].
meta-­analysis. Acta Med Acad 2015;44:135–58.
42 (CDC) CfDCaP. Seasonal influenza vaccine dosage & administration,
18 Kim MH, Rhee CK, Shim JS, et al. Inhaled corticosteroids in
2020. Available: https://www.cdc.gov/flu/about/qa/vaxadmin.htm
asthma and the risk of pneumonia. Allergy Asthma Immunol Res
[Accessed 27 Jan 2021].
2019;11:795.
43 Hancox RJ, Cowan JO, Flannery EM, et al. Bronchodilator tolerance
19 Cazeiro C, Silva C, Mayer S, et al. Inhaled corticosteroids and and rebound bronchoconstriction during regular inhaled beta-­agonist
respiratory infections in children with asthma: a meta-­analysis. treatment. Respir Med 2000;94:767–71.
Pediatrics 2017;139. 44 Harris MN, Lundien MC, Finnie DM, et al. Application of a novel
20 Torén K, Blanc PD, Qvarfordt I, et al. Inhaled corticosteroids use and socioeconomic measure using individual housing data in asthma
risk of invasive pneumococcal disease in a population-­based study. research: an exploratory study. NPJ Prim Care Respir Med
Ann Am Thorac Soc 2020;17:1570–5. 2014;24:14018.
21 Htun ZM, Aldawudi I, Katwal PC, et al. Inhaled corticosteroids as an 45 Bjur KA, Wi C-­I, Ryu E, et al. Socioeconomic status, race/ethnicity,
associated risk factor for asthmatic pneumonia: a literature review. and health disparities in children and adolescents in a mixed rural-­
Cureus 2020;12:e8717. urban community-­olmsted county, Minnesota. Mayo Clin Proc
22 Ekbom E, Quint J, Schöler L, et al. Asthma and treatment with 2019;94:44–53.
inhaled corticosteroids: associations with hospitalisations with 46 Busse WW, Morgan WJ, Taggart V, et al. Asthma outcomes
pneumonia. BMC Pulm Med 2019;19:254. workshop: overview. J Allergy Clin Immunol 2012;129:S1–8.
23 To M, To Y, Yamada H, et al. Influence of inhaled corticosteroids on 47 Fuhlbrigge A, Peden D, Apter AJ, et al. Asthma outcomes:
community-­acquired pneumonia in patients with bronchial asthma. exacerbations. J Allergy Clin Immunol 2012;129:S34–48.
Intern Med 2004;43:674–8. 48 Lenth RV. Some practical guidelines for effective sample size
24 Almirall J, Bolíbar I, Serra-­Prat M, et al. Inhaled drugs as risk factors determination. American Statistician 2001;55:187–93.
for community-­acquired pneumonia. Eur Respir J 2010;36:1080–7. 49 Akdis CA. The epithelial barrier hypothesis proposes a
doi:10.1183/09031936.00022909 comprehensive understanding of the origins of allergic and other
25 Ye Q, He X-­O, D’Urzo A. A review on the safety and efficacy of chronic noncommunicable diseases. J Allergy Clin Immunol
inhaled corticosteroids in the management of asthma. Pulmonary 2022;149:41–4.
Therapy 2017;3:1–18. doi:10.1007/s41030-017-0043-5 50 Edwards MR, Strong K, Cameron A, et al. Viral infections in allergy
26 St Sauver JL, Grossardt BR, Yawn BP, et al. Data resource profile: and immunology: how allergic inflammation influences viral infections
the Rochester epidemiology project (REP) medical records-­linkage and illness. J Allergy Clin Immunol 2017;140:909–20.
system. Int J Epidemiol 2012;41:1614–24. 51 Sykes A, Macintyre J, Edwards MR, et al. Rhinovirus-­induced
27 Dales RE, Raizenne M, el-­Saadany S, et al. Prevalence of childhood interferon production is not deficient in well controlled asthma.
asthma across Canada. Int J Epidemiol 1994;23:775–81. Thorax 2014;69:240.
28 Population of olmsted county, MN - census 2010 and 2000 52 Edwards MR, Regamey N, Vareille M, et al. Impaired innate interferon
interactive map, demographics, statistics, quick facts - induction in severe therapy resistant atopic asthmatic children.
CensusViewer. Available: http://censusviewer.com/county/MN/​ Mucosal Immunol 2013;6:797–806.
Olmsted [Accessed 7 Jan 2020]. 53 Habibzay M, Saldana JI, Goulding J, et al. Altered regulation of Toll-­
29 U.S. Census Bureau QuickFacts: Olmsted County, Minnesota. like receptor responses impairs antibacterial immunity in the allergic
Available: https://www.census.gov/quickfacts/fact/table/olmstedc​ lung. Mucosal Immunol 2012;5:524–34.
ountyminnesota/PST040218#viewtop [Accessed 5 Dec 2019]. 54 Xiao C, Puddicombe SM, Field S, et al. Defective epithelial barrier
30 Bethell CD, Kogan MD, Strickland BB, et al. A national and state function in asthma. J Allergy Clin Immunol 2011;128:549–56.
profile of leading health problems and health care quality for US 55 Vasilopoulou I, Papakonstantopoulou I, Salavoura K, et al.
children: key insurance disparities and across-­state variations. Acad Underdiagnosis and undertreatment of asthma in children: a tertiary
Pediatr 2011;11:S22–33. hospital’s experience. Clin Transl Allergy 2015;5:P19.

8 Shrestha P, et al. BMJ Open 2022;12:e051926. doi:10.1136/bmjopen-2021-051926

 Open access

56 Pando S, Lemière C, Beauchesne M-­F, et al. Suboptimal use of age children: randomised double blind placebo controlled trial. BMJ
inhaled corticosteroids in children with persistent asthma: inadequate 1997;315:858–62.
prescription, poor drug adherence, or both? Pharmacotherapy 61 Festic E, Bansal V, et al, Investigators ITGLIPS. Prehospital use of
2010;30:1109–16. inhaled corticosteroids and point prevalence of pneumonia at the
57 Taylor DR, Sears MR, Herbison GP, et al. Regular inhaled beta time of hospital admission: secondary analysis of a multicenter
agonist in asthma: effects on exacerbations and lung function. cohort study. paper presented at: Mayo clinic proceedings 2014.
Thorax 1993;48:134–8. 62 Almirall J, Bolíbar I, Serra-­Prat M, et al. New evidence of risk factors
for community-­acquired pneumonia: a population-­based study. Eur
58 Anderson HR, Ayres JG, Sturdy PM, et al. Bronchodilator treatment
Respir J 2008;31:1274–84.
and deaths from asthma: case-­control study. BMJ 2005;330:117. 63 Després F, Ducharme FM, Forget A, et al. Development and
59 Reddel HK, Busse WW, Pedersen S, et al. Should recommendations validation of a Pharmacoepidemiologic pediatric asthma control
about starting inhaled corticosteroid treatment for mild asthma be index (PPACI) using administrative data. Canadian Journal of
based on symptom frequency: a post-­hoc efficacy analysis of the Respiratory, Critical Care, and Sleep Medicine 2021;5:261–9.
START study. Lancet 2017;389:157–66. 64 Harris M, Clark J, Coote N, et al. British Thoracic Society guidelines
60 Doull IJ, Lampe FC, Smith S, et al. Effect of inhaled corticosteroids for the management of community acquired pneumonia in children:
on episodes of wheezing associated with viral infection in school update 2011. Thorax 2011;66:ii1.

Shrestha P, et al. BMJ Open 2022;12:e051926. doi:10.1136/bmjopen-2021-051926 9