ORIGINAL ARTICLE

Aspergillus Infections and Progression of Structural Lung Disease in

Children with Cystic Fibrosis
Oded Breuer1,2,3, Andre Schultz1,3,4, Luke W. Garratt1, Lidija Turkovic1, Tim Rosenow1,4, Conor P. Murray5,
Yuliya V. Karpievitch1, Lauren Akesson1,6,7, Samuel Dalton8, Peter D. Sly9, Sarath Ranganathan6,8,10,
Stephen M. Stick1,3,4*, and Daan Caudri1,3,11*; on behalf of AREST CF
1
Telethon Kids Institute and 4Division of Child Health, Faculty of Medicine and Dentistry, University of Western Australia, Perth, Western
Australia, Australia; 2Department of Pediatrics, Pediatric Pulmonary Unit, Hadassah-Hebrew University Medical Center, Jerusalem,
Israel; 3Department of Respiratory and Sleep Medicine and 5Department of Diagnostic Imaging, Perth Children’s Hospital, Perth,
Western Australia, Australia; 6Department of Paediatrics, University of Melbourne, Melbourne, Victoria, Australia; 7Victorian Clinical
Genetics Services, Murdoch Children’s Research Institute, Melbourne, Victoria, Australia; 8Department of Respiratory Medicine, Royal
Children’s Hospital, Parkville, Victoria, Australia; 9Children’s Health and Environment Program, Child Health Research Centre, The
University of Queensland, Brisbane, Queensland, Australia; 10Murdoch Children’s Research Institute, Parkville, Victoria, Australia;
and 11Department of Pediatrics/Respiratory Medicine, Erasmus Medical Center–Sophia, Rotterdam, the Netherlands
ORCID ID: 0000-0002-4775-9095 (O.B.).

Abstract Measurements and Main Results: Haemophilus influenzae,

Rationale: Recent data show that Aspergillus species are
prevalent respiratory infections in children with cystic fibrosis
(CF). The biological significance of these infections is
unknown.
Objectives: We aimed to evaluate longitudinal associations
between Aspergillus infections and lung disease in young children
with CF.
Methods: Longitudinal data on 330 children participating in
the Australian Respiratory Early Surveillance Team for Cystic
Fibrosis surveillance program between 2000 and 2018 who
underwent annual chest computed tomography (CT) imaging
and BAL were used to determine the association between
Aspergillus infections and the progression of structural lung
disease. Results were adjusted for the effects of other common
infections, associated variables, and repeated visits. Secondary
outcomes included inflammatory markers in BAL, respiratory
symptoms, and admissions for exacerbations.
Staphylococcus aureus, Pseudomonas aeruginosa, and Aspergillus
infections were all associated with worse CT scores in the same year
(Poverall , 0.05). Only P. aeruginosa and Aspergillus were associated
with progression in CT scores in the year after an infection and worse
CT scores at the end of the observation period. P. aeruginosa was
most significantly associated with development of bronchiectasis
(difference, 0.9; 95% confidence interval, 0.3–1.6; P = 0.003) and
Aspergillus with trapped air (difference, 3.2; 95% confidence interval,
1.0–5.4; P = 0.004). Aspergillus infections were also associated with
markers of neutrophilic inflammation (P , 0.001) and respiratory
admissions risk (P = 0.008).
Conclusions: Lower respiratory Aspergillus infections are
associated with the progression of structural lung disease in young
children with CF. This study highlights the need to further evaluate
early Aspergillus species infections and the feasibility, risk, and benefit
of eradication regimens.
Keywords: cystic fibrosis; Aspergillus; trapped air; bronchiectasis;
lung disease progression

( Received in original form August 14, 2019; accepted in final form November 20, 2019 )
*These authors contributed equally to this work.
AREST CF is supported by Cystic Fibrosis Foundation Therapeutics, Inc., USA and Cystic Fibrosis Australia, and National Health and Medical Research
Council grants APP1000896 and 1020555. O.B. has been supported by a Lowy Foundation Pediatric Fellowship arranged by AUSiMED (Australia/Israel
Medical Research). D.C. received support for a Research Fellowship from the Rothwell Foundation, the Ter Meulen Grant of the Royal Netherlands Academy of
Arts and Sciences, and the Sophia Childrens’ Hospital Fund. None of the funding bodies were in any way involved in the data collection, interpretation of the
data, or writing of the manuscript.
Correspondence and requests for reprints should be addressed to Oded Breuer, M.D., Telethon Kids Institute, PO Box 855, West Perth, Western Australia 6872,
Australia. E-mail: oded.breuer@telethonkids.org.au.
This article has a related editorial.
This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org.
Am J Respir Crit Care Med Vol 201, Iss 6, pp 688–696, Mar 15, 2020
Copyright © 2020 by the American Thoracic Society
Originally Published in Press as DOI: 10.1164/rccm.201908-1585OC on November 20, 2019
Internet address: www.atsjournals.org

688 American Journal of Respiratory and Critical Care Medicine Volume 201 Number 6 | March 15 2020
ORIGINAL ARTICLE
At a Glance Commentary
Scientific Knowledge on the
Subject: Recent data show that
Aspergillus species are among the most
prevalent organisms isolated from the
lower airways of young children with
cystic fibrosis. An association between
Aspergillus infection and worse clinical
outcomes, such as increased rate of
respiratory exacerbation, low lung
function, and worse structural lung
disease, has been previously reported.
However, in view of lack of sufficient
longitudinal data, it has been
suggested that Aspergillus infections
may represent an epiphenomenon and
are not causally related to lung disease.
What This Study Adds to the Field:
Data from this longitudinal
observational cohort indicate that
lower respiratory tract infections with
Aspergillus species during preschool
age are associated with progressive
structural lung disease. Children
infected with Aspergillus species
experienced worse lung disease at the
time of infection and a more rapid
progression of structural lung disease
in the years after the infection. This
was highly associated with the number
of Aspergillus infections experienced in
the first 5 years and persisted after
adjusting for multiple variables
associated with underlying disease
severity. Further investigation is
required to evaluate the feasibility and
safety of eradication protocols for
early lower respiratory Aspergillus
species infections.
In infants and young children with cystic
fibrosis (CF), interactions between lower
respiratory infections and the associated
inflammatory response have been shown to
drive the progression of structural lung
disease and lung function decline (1–6).
Staphylococcus aureus, Haemophilus
influenzae, and Pseudomonas aeruginosa
were considered to be the most prevalent
respiratory infections in these young
Author Contributions: Study conception and design: O.B., S.M.S., and D.C. Data acquisition: O.B., T.R., C.P.M., S.D., D.C., and AREST CF. Data analysis:
O.B., L.T., Y.V.K., and D.C. Data interpretation: O.B., A.S., L.W.G., P.D.S., S.R., S.M.S., and D.C. Manuscript draft: O.B. Manuscript revision: all authors,
including AREST CF. Approval of final version of the manuscript: all authors, including AREST CF. Accountability for all aspects of the research: all authors,
including AREST CF.
Breuer, Schultz, Garratt, et al.: Aspergillus and Lung Disease in Children with CF
children (7–11). However, recent data show
an increase in the relative prevalence of
Aspergillus species (10, 12–15), once
considered a late-occurring infection
(16–18). Moreover, the decline in
prevalence of common early CF-associated
pathogens has resulted in Aspergillus now
being one of the most prevalent organisms
isolated from the lower airways of young
children with CF (13–15, 19).
Despite this relative dominance of
Aspergillus infections in young children
with CF, there is currently no consensus on
the clinical significance of lower airway
Aspergillus infection or the need for routine
screening and treatment in the absence of
allergic bronchopulmonary aspergillosis
(ABPA).
Studies have shown a role for CFTR in
mediating epithelial cell responses to
Aspergillus fumigatus and that Aspergillus
infection is associated with increased
airway inflammation, suggesting a
pathogenic role for Aspergillus in patients
with CF (4, 20, 21). An association between
Aspergillus infection and worse clinical
outcomes, such as increased rate of
respiratory exacerbation, low or
deteriorating lung function, and worse
structural lung disease, have been
previously reported (13, 15, 20, 22–25).
However, these results are inconsistent and
are largely based on cross-sectional data
(15, 24, 26, 27).
The aim of the current study was to
assess longitudinal associations between
lower airway infections with Aspergillus
species and current as well as future
indicators of lung disease severity in
preschool children with CF. Some of the
results of these studies have been previously
reported in the form of an abstract (28).
Methods
Study Population
All children participating in the AREST CF
(Australian Respiratory Early Surveillance
Team for Cystic Fibrosis) surveillance
program at Princess Margaret Hospital and
at the Royal Children’s Hospital between
the January 2000 and February 2018 for
whom at least one chest computed
tomography (CT) scan was available were
included in the study. A more detailed
explanation of the methods is provided in
the online supplement. The AREST CF
prospective observational cohort study
includes detailed clinical and laboratory
assessments at 3 months of age and yearly
thereafter until 6 years of age, as previously
described (29). These assessments include a
bronchoscopy with BAL performed under
general anesthesia when clinically stable
and, since 2005, a pressure-controlled chest
CT scan performed just before the
bronchoscopy. Written consent was
obtained from parents before each visit
(Child and Adolescent Health Service
Human Research Ethics Committee,
EP1972).
Bronchoscopy, BAL Collection, and
Processing
Detailed description of BAL collection,
microbiological processing, and analysis of
inflammatory markers was previously
described (4, 29, 30). BAL was performed
after the induction of general anesthesia
and chest CT scan, using a flexible
bronchoscope through a laryngeal mask
airway. BAL samples were analyzed for
markers of inflammation (cell count with
differential, free neutrophil elastase activity,
and IL-8 concentration) and standard
microbiological processing. Specifically, for
the detection of yeast and fungi, 20-ml BAL
samples were inoculated on Sabouraud agar
with chloramphenicol (Sigma-Aldrich) and
incubated aerobically for 48 hours at 358 C
in 5% CO2. Cultures were read at 24 and 48
hours and further incubated for 12 days in
air at 288 C to observe for fungal growth.
Any growth was considered positive.
Structural Lung Disease
CT scans consisted of limited slice (before
2007) or volumetric (since 2007) scans
obtained at both end inspiration and end
expiration (29). Chest CT scans were
scored using the CF-CT and the
Perth–Rotterdam Annotated Grid
Morphometric Analysis (PRAGMA-CF)
scoring methods (31, 32). Both systems
were used to assess our main outcome of
structural lung disease.
689
 ORIGINAL ARTICLE

Table 1. Study Population Characteristics

Characteristic All (Unique Children) Aspergillus Ever Positive* Aspergillus Negative* P Value†

No. of study participants 330 115 215 —

Sex, M, n (%) 162 (49) 57 (50) 105 (49) 0.900
Pancreatic insufficiency 85 (280/330) 85 (98/115) 66 (142/215) 0.090
Meconium ileus 19 (62/330) 25 (29/115) 15 (33/215) 0.072
Homozygous DF508 50 (164/330) 54 (61/114) 49 (103/211) 0.419

Aspergillus-Positive Aspergillus-Negative
All (Visits) Sample Sample P Value‡

No. of visits with BAL sample 1,631 186 1,445 —

Age at visit, yr, median (IQR) 3.0 (1.2–4.5) 3.9 (2.3–5.1) 2.9 (1.2–4.3) <0.001
P. aeruginosa infection 9.5 (130/1,631) 11.8 (22/186) 7.5 (108/1,445) 0.165
Tobramycin inhalation use in the 3 mo before BAL 9.7 (127/1,315) 16.6 (27/163) 8.7 (100/1,152) 0.006
Tobramycin inhalation use in the year before BAL 16 (225/1,315) 21 (36/168) 16 (189/1,213) 0.084
Azithromycin use in the 3 mo before BAL 6.0 (79/1,315) 11.0 (18/163) 5.3 (61/1,152) 0.030
Oral amoxicillin clavulanate use in the 3 mo 54 (704/1,315) 58 (95/163) 53 (609/1,152) 0.144
before BAL
Dornase alfa inhalation in the 3 mo before BAL 10 1 (32/1,315) 17 (28/163) 10 (109/1,043) 0.001
i.v. antibiotic use in year before BAL 22.5 (228/1,013) 35.1 (46/131) 20.6 (182/882) 0.002

Definition of abbreviations: IQR = interquartile range; P. aeruginosa = Pseudomonas aeruginosa.

Data are presented as % (n/N) unless otherwise noted. Bold indicates P , 0.05.
*Ever having a positive BAL culture versus not having a single BAL culture positive for Aspergillus during the entire study period.
†
Comparison of children ever positive versus children never positive by means of Student’s t test (continuous variable) or chi-square test (dichotomous
variables).
‡
Comparison between visits with Aspergillus-positive and Aspergillus-negative BAL cultures, using univariate mixed-effects logistic regression models
taking into account repeated visits in the same children.

Clinical Outcomes
Clinical outcomes included: respiratory
symptoms at the time of the bronchoscopy
(cough or wheeze), the number of severe
pulmonary exacerbations (number of
intravenous antibiotic treatment courses)
in the year after the annual assessment,
and lung function at 6 years of age.
Statistical Analysis
Children’s characteristics, age, pancreatic
insufficiency, homozygosity for CFTR
p.(Phe508del), medication use, and other
infections at baseline were evaluated as risk
factors for Aspergillus infection at any age
using mixed effect logistic regression.
Multivariate linear regression models
were used to evaluate the independent
associations of common lower respiratory
infections detected on BAL with structural
lung disease and clinical outcomes.
Analyses were additionally adjusted for
pancreatic insufficiency, age, sex, and
homozygosity for CFTR p.(Phe508del).
Sensitivity analyses were performed to
evaluate the relative effect of suspected
ABPA. All analyses were longitudinal and
used mixed-effects models to account for
repeated visits. Aspergillus species infections
were associated first with CT subscores
obtained during the same annual
assessment (same year results) and second
with CT subscores at the following annual
assessment (1-yr progression results),
adjusted also for baseline CT scores. Third,
the association between Aspergillus species
infection during the first 5 years of life was
related to CT subscores at end of follow-up,
defined as age 6 years or, if unavailable, age
5 years (end-of-study results). Additional
subanalyses were performed to evaluate the
effects of possible confounders related to
worse respiratory disease to account for the
possibly that Aspergillus is a marker of poor
lung disease and not a causative pathogenic
infection (see online supplement). All
analyses were performed using Stata
version 15 (StataCorp).
Results
Study Population
Data from 330 children participating in the
AREST CF cohort were evaluated in this
study; the clinical characteristics of these
children are presented in Table 1. The
included children underwent a total of
1,631 annual assessments with BALs. From
these assessments, 186 BALs (11.4%) from
115 children (35%) cultured positive for
Aspergillus species infection: 75%
Aspergillus fumigatus, 13% Aspergillus
niger, 4% Aspergillus terreus, 2% Aspergillus
flavus, and 12% unidentified Aspergillus
species. Six percent of positive cultures
identified more than one Aspergillus type.
Twenty-two out of the 186 cultures
positive for Aspergillus were coinfected
with P. aeruginosa (11.8%). The risk
for a coinfection of Aspergillus with
P. aeruginosa was not significantly different
from the risk of a coinfection with
P. aeruginosa for any of the other commonly
cultured lower respiratory infections
(H. influenzae, 22 out of 160 [14%] and
S. aureus, 26 out of 187 [14%]; P = 0.8)
(see Table E1 in the online supplement:
infection and coinfection prevalence).
The most common pathogens isolated
are presented in Table E1. Other isolates
included Streptococcus pneumoniae (60
positive BALs [3.7%] in 52 children [16%]),
Stenotrophomonas maltophilia (48 positive
BALs [3.5%] in 39 children [12%]), and
Moraxella catarrhalis (29 positive BALs
[1.8%] in 27 children [8%]).
Available for analysis of the main
outcome of structural lung disease in the 330
included children were 1,035 chest CT scans

690 American Journal of Respiratory and Critical Care Medicine Volume 201 Number 6 | March 15 2020
ORIGINAL ARTICLE

structural lung disease at the end of follow-

Age
up (Table 2, end-of-study results). This
Pancreatic insufficiency outcome showed a dose–response
relationship with an incremental worsening
P. aeruginosa of end-of-study lung disease for each
additional time Aspergillus was cultured
H. influenza during the first 5 years of life (Figures 3 and
S. aureus
E1). The age of first Aspergillus infection
was not an independent risk factor for
S. maltophilia worse structural lung disease.
Although all four common pathogens
Tobramycin use were associated with structural lung disease
in the same year as BAL, Pseudomonas and
Azithromicine use
Aspergillus species showed the strongest
Treatment with i.v. antibiotics effects and were the only two pathogens
associated with future progression of
Dornase alfa use bronchiectasis and overall structural lung
disease. Aspergillus species infections were
Hypertonic saline use most consistently and strongly associated
with the development of worse trapped air
0.17 0.25 0.5 1.0 2.0 4.0 6.0
(mosaic attenuation) and mucus plugging,
Odds ratios with 95% confidence intervals
whereas P. aeruginosa infections were more
Figure 1. Odds ratio for risk of Aspergillus species infection. Results are from mixed-effects logistic consistently associated with the presence and
regression models clustered for repeated visits in the same child. Odds ratios of the mixed-effects further progression of bronchiectasis. In the
model are: age, 1.3; 95% confidence interval (CI), 1.1–1.5 (P = 0.001); pancreatic insufficiency, 3.0; subgroup of children (n = 260 children, 649
95% CI, 1.1–8.3 (P = 0.036); Pseudomonas aeruginosa, 1.5; 95% CI, 0.7–3.1 (P = 0.300);
CT scans) with both CF-CT and PRAGMA-
Haemophilus influenzae, 1.3; 95% CI, 0.6–2.6 (P = 0.475); Staphylococcus aureus, 1.1; 95% CI,
0.5–2.2 (P = 0.860); Stenotrophomonas maltophilia, 1.2; 95% CI, 0.4–3.6 (P = 0.860); tobramycin use
CT scores available, the results of these
in prior 3 months, 2.3; 95% CI, 1.1–5.1 (P = 0.033); azithromycin use in prior 3 months, 1.1; 95% CI, analyses remained significant, with a strong
0.4–2.9 (P = 0.851); intravenous antibiotic treatment in prior year, 1.4; 95% CI, 0.8–2.5 (P = 0.244); association between trapped air scores
dornase alfa use in prior 3 months, 1.4; 95% CI, 0.6–3.3 (P = 0.456); and hypertonic saline use in prior (P , 0.001) and bronchiectasis scores
3 months, 1.0; 95% CI, 0.4–2.8 (P = 0.970). (P , 0.001) for both methods.
To further evaluate if Aspergillus is
independently associated with lung disease,
in 310 children scored using the CF-CT recent use of inhaled tobramycin (OR, 2.3; we repeated analyses with adjustment for an
scoring method and 745 scans in 276 children 95% CI, 1.1–5.0; P = 0.036) remained even wider range of variables possibly
scored using the PRAGMA-CF method, of significant and independent predictors indicating worse lung disease (described in
which 649 scans (87%) in 260 children (94%) for Aspergillus (Figure 1). None of the other the online supplement under Statistical
were scored by both methods. A total of 179 respiratory pathogens were associated with an Analyses), noting this considerably reduced
children had a CT scan at the end of follow-up increased risk for Aspergillus species sample size with complete data available. The
at 5 or 6 years; median CT scores at the end of infections. Treatment with intravenous association of Aspergillus species infections
follow-up are presented in Table E2. antibiotics for respiratory exacerbations in the with trapped air and mucus plugging scores
Data regarding routine treatment were year before BAL was not an independent risk remained highly significant for the same year,
available for 326 patients (98%) in 1,381 factor for Aspergillus species infections. the 1-year progression, and the end-of-study
annual assessments (85%). Only three analyses (Table E3), despite reduction in
patients reported to have received antifungal Association of Aspergillus Species power and comprehensive adjustments.
therapy during the study period. with the Presence and Progression of Total IgE levels and Aspergillus-
Structural Lung Disease specific IgE levels were available for
Factors Associated with Aspergillus Children with positive Aspergillus species analyses in 70% of visits for children
Species Infection on BAL cultures had significantly worse who were Aspergillus positive. During the
Table 1 (univariable analysis) shows that current structural lung disease (Table 2, study period, only four children (at six
treatments with oral, inhaled, and same-year results) and a significant visits) had total IgE levels . 500 IU/ml
intravenous antibiotics were all significant progression in their structural lung disease with positive Aspergillus-specific IgE
risk factors for Aspergillus species infection. in the following year (Table 2, 1-yr levels, thereby qualifying as suspected
However, when combining all significant progression results). Furthermore, children ABPA. A sensitivity analysis excluding
univariate factors in a multivariable who ever had Aspergillus species cultured these children did not affect our results,
adjusted model, only age (odds ratio [OR], from their BAL during the first 5 years of making it unlikely that the associations
1.3; 95% confidence interval [CI], 1.1–1.5; life experienced a more rapid progression in between Aspergillus and structural lung
P , 0.001), pancreatic insufficiency their structural lung disease during those disease in this age group can be explained
(OR, 3.2; 95% CI, 1.1–8.9; P = 0.026), and years (Figure 2), resulting in worse by ABPA (Table E4).

Breuer, Schultz, Garratt, et al.: Aspergillus and Lung Disease in Children with CF 691
 ORIGINAL ARTICLE

Table 2. Association of Structural Lung Disease on Chest CT Scans with Pathogens Detected on BAL

P. aeruginosa Aspergillus Species H. influenzae S. aureus

Same-year results (at the

time of bronchoscopy)*
CF-CT score (1,030 visits,
307 children)
Bronchiectasis 1.3 (0.8 to 1.8), <0.001 0.7 (0.2 to 1.1), 0.003 0.8 (0.3 to 1.3), 0.001 0.6 (0.1 to 1.0), 0.021
Bronchial wall 1.1 (0.4 to 1.9), 0.003 1.0 (0.4 to 1.7), 0.001 1.0 (0.4 to 1.8), 0.001 0.3 (20.4 to 1.0), 0.370
thickening
Trapped air 0.9 (0.3 to 1.4), 0.003 1.6 (1.2 to 2.1), <0.001 0.6 (0.1 to 1.1), 0.026 0.6 (0.1 to 1.1), 0.022
Mucus plugging 0.4 (0.2 to 0.6), <0.001 0.5 (0.3 to 0.7), <0.001 0.1 (20.1 to 0.3), 0.441 0.3 (0.1 to 0.4), 0.010
PRAGMA score (741
visits, 272 children)
Bronchiectasis, % 0.7 (0.4 to 0.9), <0.001 0.3 (0.03 to 0.5), 0.026 0.1 (20.2 to 0.3), 0.573 20.2 (20.4 to 0.1), 0.218
Trapped air, % 3.6 (2.0 to 5.2), <0.001 5.3 (4.0 to 6.5), <0.001 1.3 (20.3 to 2.9), 0.106 0.8 (20.7 to 2.3), 0.298
Disease, % 1.4 (0.9 to 1.9), <0.001 1.0 (0.6 to 1.4), <0.001 0.2 (20.3 to 0.6), 0.485 0.01 (20.4 to 0.5), 0.950
1-year progression results
(at 1 yr after the
bronchoscopy)*†
CF-CT score (634 visits,
222 children)
Bronchiectasis 0.4 (20.3 to 1.1), 0.273 0.4 (20.2 to 1.0), 0.217 20.4 (21.1 to 0.3), 0.312 0.4 (20.2 to 1.1), 0.198
Bronchial wall 0.2 (20.7 to 1.1), 0.675 0.9 (0.1 to 1.7), 0.030 20.1 (21.1 to 0.9), 0.848 20.4 (21.3 to 0.5), 0.378
thickening
Trapped air 20.4 (21.1 to 0.3), 0.304 1.1 (0.4 to 1.7), 0.001 0.1 (20.7 to 0.8), 0.845 0.6 (20.1 to 1.3), 0.077
Mucus plugging 0.3 (20.1 to 0.6), 0.056 0.3 (0.1 to 0.6), 0.013 20.1 (20.4 to 0.2), 0.403 0.2 (20.1 to 0.4), 0.192
PRAGMA score (406
visits, 143 children)
Bronchiectasis, % 0.5 (0.1 to 0.9), 0.020 0.6 (0.2 to 0.9), 0.001 20.3 (20.7 to 0.2), 0.221 0.1 (20.3 to 0.6), 0.636
Trapped air, % 20.4 (22.9 to 2.1), 0.762 3.4 (1.3 to 5.6), 0.002 22.1 (24.9 to 0.7), 0.140 0.8 (21.9 to 3.5), 0.562
Disease, % 0.7 (20.1 to 1.4), 0.069 0.7 (0.1 to 1.3), 0.029 20.3 (21.1 to 0.5), 0.491 20.4 (21.2 to 0.4), 0.330
End-of-study results (at 5 or
6 yr of age)‡
CF-CT score (179
children)
Bronchiectasis 1.6 (0.4 to 2.8), 0.009 1.0 (20.2 to 2.2), 0.089 20.4 (21.6 to 0.8), 0.476 1.2 (20.3 to 2.3), 0.044
Bronchial wall 1.0 (20.1 to 2.0), 0.071 1.6 (0.6 to 2.6), 0.002 0.3 (20.7 to 1.4), 0.534 0.5 (20.5 to 1.6), 0.334
thickening
Trapped air 1.4 (0.5 to 2.3), 0.003 1.0 (0.5 to 1.9), 0.034 20.6 (21.5 to 0.3), 0.205 0.5 (20.4 to 1.5), 0.238
Mucus plugging 0.6 (0.1 to 1.2), 0.014 0.5 (0.1 to 1.0), 0.031 20.3 (20.8 to 0.2), 0.279 0.4 (20.1 to 1.0), 0.081
PRAGMA score (161
children)
Bronchiectasis, % 0.8 (0.3 to 1.4), 0.005 0.6 (0.03 to 1.2), 0.040 20.4 (21.0 to 0.2), 0.208 0.2 (20.4 to 0.8), 0.456
Trapped air, % 3.0 (0.9 to 5.1), 0.005 2.9 (0.8 to 4.9), 0.007 0.1 (22.0 to 2.3), 0.892 1.0 (21.1 to 3.2), 0.328
Disease, % 1.0 (0.3 to 1.7), 0.007 1.2 (0.4 to 1.9), 0.002 20.5 (21.3 to 0.2), 0.181 0.1 (20.6 to 0.9), 0.698

Definition of abbreviations: CF = cystic fibrosis; CT = computed tomography; H. influenzae = Haemophilus influenzae; P. aeruginosa = Pseudomonas
aeruginosa; PRAGMA = Perth-Rotterdam Annotated Grid Morphometric Analysis; S. aureus = Staphylococcus aureus.
Data are presented as mean difference if pathogen present compared with pathogen absent (95% confidence interval), P value. Bold indicates P , 0.05.
*Results from mixed-effects models. Analyses were clustered for repeated measures in the same child and adjusted for pancreatic insufficiency, age, sex,
homozygosity for DF508, and other infections.
†
Additionally adjusted for baseline CT score results at time of airway sampling (bronchoscopy).
‡
Results from linear regression models for children ever infected in the first 5 years of life and assessed by a CT scan at their last available study
appointment (5 or 6 yr of age). Analysis was adjusted for pancreatic insufficiency, sex, homozygosity for DF508, and other infections.

Clinical Outcome Measures
BALs positive for S. aureus, H. influenzae,
P. aeruginosa, and Aspergillus species
infections were all significantly associated
with markers of neutrophilic inflammation
on BAL (Table 3). There were no significant
differences in inflammatory marker levels
between these infections (IL-8, P = 0.202;
neutrophil elastase, P = 0.072; percentage
neutrophil on BAL, P = 0.587).
Children with Aspergillus species
infection had significantly increased
number of intravenous antibiotic courses in
the year after the infection (difference of 0.2
admissions per year; 95% CI, 0.04–0.3;
P = 0.013). This association also remained
significant after controlling for previous
inpatient and outpatient intravenous
antibiotic treatment courses. Children in
whom Aspergillus species was cultured from
their BAL were more likely to report
respiratory symptoms (cough or wheeze) at
the time of the bronchoscopy (OR, 1.7; 95%
CI, 1.0–3.0; P = 0.051).

692 American Journal of Respiratory and Critical Care Medicine Volume 201 Number 6 | March 15 2020
ORIGINAL ARTICLE

PsA ASP HFLU MSSA

6 6 6 6
Bronchiectasis

4 4 4 4
score

2 2 2 2

0 0 0 0
0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6

5 5 5 5
Trapped air
score

3 3 3 3

1 1 1 1
0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6

1.5 1.5 1.5 1.5

Mucus plugging

1 1 1 1
score

.5 .5 .5 .5

0 0 0
0

0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6

10 10 10 10
thickening score
Bronchial wall

8 8 8 8
6 6 6 6
4 4 4 4
2 2 2 2
0 0 0 0
0 2 4 6 0 2 4 6 0 2 4 6 0 2 4 6
Age
Ever infected Never infected 95% CI
Figure 2. Progression of structural lung disease (CF-CT subscores) during the first 6 years of life in children ever infected versus never infected with a
specific infection (one of the four most commonly cultured lower respiratory tract infections). Actual data of computed tomography subscores (95%
confidence interval [CI]), unadjusted for associated variables, are plotted as a function of age at study visit using a fractional polynomial model
(twoway fpfitci procedure in Stata). ASP = Aspergillus; CF = cystic fibrosis; CT = computed tomography; HFLU = Haemophilus influenzae;
MSSA = Staphylococcus aureus; PsA = Pseudomonas aeruginosa.

Discussion
In this unique longitudinal study, early
Aspergillus species infections increase the
risk for the presence and increased
progression of structural lung disease,
predominantly characterized by trapped
air and mucus plugging. This association
was significant when assessing cross-
sectional as well as short-term and long-
term longitudinal effects, suggesting a
detrimental role for Aspergillus infection on
CF airways. This study suggests a need for
addressing early Aspergillus infections,
commencing with the development of
safe, effective eradication therapies.
Our results are consistent with previous
studies showing a strong association
between Aspergillus infection and increased
trapped air, presence of bronchiectasis, and
increase in total disease score on chest CT
scans in the same year of infection (15, 20,
24). These studies show cross-sectional
findings, which may reflect confounding by
disease severity and were unable to describe
the long-term consequences of Aspergillus
infections. To overcome this limitation, our
study analyzed longitudinal data from
preschool-aged children with an overall
mild structural lung disease at baseline,
relatively naive to previous infections, while
controlling for other airway pathogens and
baseline variables known to be associated
with worse pulmonary disease. Our data
provide important, new, clinically relevant
insights. First, current Aspergillus lower
respiratory infections are associated with a
significantly increased progression in
structural lung disease within 1 year after
the infection. Second, children ever infected
with Aspergillus in the first 5 years of life
experience a more rapid progression in
their structural lung disease during
childhood, resulting in worse end-of-study
chest CT outcomes. Third, the severity of
the resulting structural lung disease is
highly associated with the number of
Aspergillus infections experienced in the

Breuer, Schultz, Garratt, et al.: Aspergillus and Lung Disease in Children with CF 693
 ORIGINAL ARTICLE

% Bronchiectasis at end of study

12 40

% Trapped air at end of study

% Disease at end of study

10

30 8
8
6
20
4
4
10
2

0 0 0
0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5
Number of previous Aspergillus infections Number of previous Aspergillus infections Number of previous Aspergillus infections
Figure 3. Computed tomography (CT) subscores at end of follow-up (end-of-study chest CT Perth–Rotterdam Annotated Grid Morphometric
Analysis–Cystic Fibrosis scores) according to the number of times Aspergillus species was cultured in BALs during the first 5 years of life. Actual
unadjusted data of CT subscores (95% confidence interval) are plotted as a function of the number of Aspergillus infections in the first 5 years of life using a
fractional polynomial model (twoway fpfitci procedure in Stata).

first 5 years of life, suggesting that the 27). Indeed, we found the risk of Aspergillus analyses performed, Aspergillus infections
greatest impact on long-term outcomes is infections to be related to variables remained strongly and significantly
likely to be through early eradication and representing worse lung disease (pancreatic associated with current and future
prevention of persistent Aspergillus insufficiency, increasing age, more intensive structural lung disease.
infections. treatment), in line with previous reports The only other infection that was found
Previous studies have raised the (26, 33, 34). However, adjustment for the to be associated with a similar progression in
question of whether Aspergillus infections above and additional possible confounding structural lung disease and worse end-of-study
represent an epiphenomenon that is not and intermediate factors did not change our outcomes was P. aeruginosa. We did not
causally related to lung disease (15, 19, 26, results and interpretations. Throughout all find a significant interaction between

Table 3. Association of Clinical Outcomes with Pathogens Detected on BAL

P. aeruginosa Aspergillus Species H. influenzae S. aureus

Neutrophilic inflammation
on BAL (1,565 visits,
325 children)
IL-8 levels (log scale) 0.7 (0.5 to 1.0), <0.001 0.5 (0.3 to 0.7), <0.001 0.6 (0.3 to 0.8), <0.001 0.3 (0.1 to 0.5), 0.008
Percent neutrophil 9.8 (6.0 to 13.7), <0.001 8.2 (4.8 to 11.6), <0.001 5.9 (2.3 to 9.6), 0.001 11.9 (8.6 to 15.3), <0.001
count
Neutrophil elastase 0.7 (0.5 to 0.9), <0.001 0.5 (0.3 to 0.7), <0.001 0.4 (0.2 to 0.6), 0.001 0.3 (0.1 to 0.5), 0.013
level (log scale)
No. of admissions for 0.5 (0.4 to 0.7), <0.001 0.2 (0.1 to 0.3), 0.008 20.1 (20.3 to 0.01), 0.078 0.1 (20.1 to 0.2), 0.069
i.v. antibiotic
therapy* (793 visits,
225 children)
FEV1 at 6 yr of age† 0.2 (20.3 to 0.6), 0.475 0.3 (20.1 to 0.7), 0.123 20.3 (20.7 to 0.1), 0.160 20.1 (20.5 to 0.3), 0.733
(155 children)
BMI at 6 yr of age† (146 0.5 (20.1 to 1.1), 0.098 20.2 (20.8 to 0.4), 0.490 20.01 (20.6 to 0.6), 0.997 20.7 (21.3 to 0.1), 0.026
children)
Respiratory symptoms 1.4 (0.7 to 2.8), 0.239 1.7 (1.0 to 3.0), 0.051 1.5 (0.8 to 2.7), 0.223 1.8 (1.0 to 3.3), 0.050
at the time of the
bronchoscopy
(cough or wheeze)
(731 visits, 278
children), OR (95%
CI), P value

Definition of abbreviations: BMI = body mass index; CI = confidence interval; H. influenzae = Haemophilus influenzae; OR = odds ratio; P. aeruginosa =
Pseudomonas aeruginosa; S. aureus = Staphylococcus aureus.
Data are presented as difference (95% confidence interval), P value, unless otherwise noted. Results from mixed-effects models were clustered for
repeated measures in same child. Bold indicates P , 0.05.
*In the 12 months after the BAL (additionally adjusted for the number of intravenous antibiotic courses in the year before BAL).
†
FEV1 and BMI were used at a single time point at end of follow-up (age 6 yr) and analyzed with linear regression models. All analyses are adjusted for
pancreatic insufficiency, age, sex, homozygosity for DF508, and other infections as detected from BAL in the same year.

694 American Journal of Respiratory and Critical Care Medicine Volume 201 Number 6 | March 15 2020
ORIGINAL ARTICLE

Aspergillus and P. aeruginosa infections
(as predictors of future infections or as
coinfections), and both infections were
independently associated with structural
lung disease progression. Importantly, the
nature of structural lung disease
progression was distinct between these two
infections. Both infections were associated
with the development of worse overall
disease and bronchiectasis. However,
Aspergillus infections were observed to be
most strongly associated with the presence
and progression of trapped air and mucus
plugging, whereas P. aeruginosa primarily
and most significantly associated with
bronchiectasis. The underlying physiology
of trapped air (mosaic attenuation) is not
well understood, but it has been previously
shown to be associated with increased rate
of respiratory exacerbation, the
development of bronchiectasis, and worse
lung function (present and future) (6, 35,
36). This unique association between
Aspergillus infections and increased trapped
air on chest CT scans has been described
earlier (15, 24) and may support a specific
pathogenic role for Aspergillus infections on
underlying disease biology, different from
that of pathogenic bacteria. Indeed, it has
been suggested that CF airways have an
innate susceptibility to Aspergillus species
infections, resulting in an exaggerated
inflammatory response and upregulated
mucus production, which may explain our
findings (4, 20, 21, 37).
Our study also demonstrated an
association between current Aspergillus
species infections and other clinical
outcomes. Lower respiratory infections
with Aspergillus species were associated
with respiratory symptoms and an
increased risk of future intravenous
antibiotic treatment, in line with previous
reports (22, 25).
The strengths of our study lie in its
prospective longitudinal design, number of
children with regular assessments, and
uniform sampling and culturing techniques
of lower airway samples over nearly two
decades. Notably, in the current study we
analyzed two different chest CT scoring
systems: the CF-CT score, which in young
children with mild lung disease is mostly a
qualitative score, and the PRAGMA-CF
score, a quantitative score. In both scoring
systems, current Aspergillus infections
were highly significantly associated with
current disease, 1-year progression, and
worse end-of-study chest CT scores,
further substantiating the validity and
robustness of our findings. Available
evidence suggests that structural lung
disease in pediatric patients with CF is
a good predictor of future outcomes
and lung function results, indicating
the clinical relevance of our findings
(38–40).
As these are observational data, the risk
of residual confounding and/or reverse
causation remains. Other potential
limitations include possible mutual
inhibitory effects of Aspergillus and
P. aeruginosa on in vitro culture growth,
which might have affected the true
prevalence of infections with these
organisms (19). Furthermore, treatment of
older patients with the new CFTR
modulator therapies was recently also
shown to influence the prevalence of
Aspergillus infections (41).
Nevertheless, our observations fulfill
classical criteria for causation (42),
including: the strength, clear temporality,
and dose–response effect of the association;
the specificity of associated damage
(trapped air); the consistency of the
association in our different subanalyses
as well as with previous studies (15, 20);
and the biological plausibility of harmful
effects of Aspergillus infections. Although
causality cannot be proven, currently
available evidence supports a role for
Aspergillus infections in the development
of early structural lung disease in children
with CF.
In summary, Aspergillus species have
become one of the most commonly
cultured pathogens from the lower
respiratory tracts of adults and children
with CF. Results from this study suggest
that early Aspergillus infections are an
independent risk factor for increased
progression of lung disease in young
children with CF. These findings highlight a
need for routine screening of Aspergillus
infections and the need to evaluate the
feasibility and risk versus benefit of
eradication protocols. n
Author disclosures are available with the text
of this article at www.atsjournals.org.

References
1. Rosenfeld M, Gibson RL, McNamara S, Emerson J, Burns JL, Castile R,
et al. Early pulmonary infection, inflammation, and clinical outcomes in
infants with cystic fibrosis. Pediatr Pulmonol 2001;32:356–366.
2. Sagel SD, Kapsner R, Osberg I, Sontag MK, Accurso FJ. Airway
inflammation in children with cystic fibrosis and healthy children
assessed by sputum induction. Am J Respir Crit Care Med 2001;164:
1425–1431.
3. Sagel SD, Gibson RL, Emerson J, McNamara S, Burns JL, Wagener JS,
et al.; Inhaled Tobramycin in Young Children Study Group; Cystic
Fibrosis Foundation Therapeutics Development Network. Impact of
Pseudomonas and Staphylococcus infection on inflammation and
clinical status in young children with cystic fibrosis. J Pediatr 2009;
154:183–188.
4. Gangell C, Gard S, Douglas T, Park J, de Klerk N, Keil T, et al.; AREST
CF. Inflammatory responses to individual microorganisms in the lungs
of children with cystic fibrosis. Clin Infect Dis 2011;53:425–432.
5. Pillarisetti N, Williamson E, Linnane B, Skoric B, Robertson CF, Robinson
P, et al.; Australian Respiratory Early Surveillance Team for Cystic
Fibrosis (AREST CF). Infection, inflammation, and lung function
decline in infants with cystic fibrosis. Am J Respir Crit Care Med 2011;
184:75–81.
6. Sly PD, Gangell CL, Chen L, Ware RS, Ranganathan S, Mott LS,
et al.; AREST CF Investigators. Risk factors for bronchiectasis
in children with cystic fibrosis. N Engl J Med 2013;368:
1963–1970.
7. Burns JL, Gibson RL, McNamara S, Yim D, Emerson J, Rosenfeld M,
et al. Longitudinal assessment of Pseudomonas aeruginosa in
young children with cystic fibrosis. J Infect Dis 2001;183:
444–452.
8. West SE, Zeng L, Lee BL, Kosorok MR, Laxova A, Rock MJ, et al.
Respiratory infections with Pseudomonas aeruginosa in children with
cystic fibrosis: early detection by serology and assessment of risk
factors. JAMA 2002;287:2958–2967.
9. Cystic Fibrosis Foundation Patient Registry annual data report 2016.
Bethesda, MD: Cystic Fibrosis Foundation; 2016.
10. Australian Cystic Fibrosis Data Registry annual report 2015. The
Australian Cystic Fibrosis Data Registry; 2015.
11. Heltshe SL, Khan U, Beckett V, Baines A, Emerson J, Sanders DB, et al.
Longitudinal development of initial, chronic and mucoid
Pseudomonas aeruginosa infection in young children with cystic
fibrosis. J Cyst Fibros 2018;17:341–347.
12. Ramsey KA, Hart E, Turkovic L, Padros-Goossens M, Stick SM,
Ranganathan SC. Respiratory infection rates differ between
geographically distant paediatric cystic fibrosis cohorts. ERJ Open
Res 2016;2:00014-2016.

Breuer, Schultz, Garratt, et al.: Aspergillus and Lung Disease in Children with CF 695

13. Saunders RV, Modha DE, Claydon A, Gaillard EA. Chronic Aspergillus
fumigatus colonization of the pediatric cystic fibrosis airway is
common and may be associated with a more rapid decline in lung
function. Med Mycol 2016;54:537–543.
14. Breuer O, Schultz A, Turkovic L, de Klerk N, Keil AD, Brennan S, et al.
Changing prevalence of lower airway infections in young children
with cystic fibrosis. Am J Respir Crit Care Med 2019;200:590–599.
15. Harun SN, Wainwright CE, Grimwood K, Hennig S; Australasian Cystic
Fibrosis Bronchoalveolar Lavage (ACFBAL) study group. Aspergillus
and progression of lung disease in children with cystic fibrosis.
Thorax 2019;74:125–131.
16. Lipuma JJ. The changing microbial epidemiology in cystic fibrosis. Clin
Microbiol Rev 2010;23:299–323.
17. Gibson RL, Burns JL, Ramsey BW. Pathophysiology and management
of pulmonary infections in cystic fibrosis. Am J Respir Crit Care Med
2003;168:918–951.
18. Bargon J, Dauletbaev N, Köhler B, Wolf M, Posselt HG, Wagner TO.
Prophylactic antibiotic therapy is associated with an increased
prevalence of Aspergillus colonization in adult cystic fibrosis
patients. Respir Med 1999;93:835–838.
19. Liu JC, Modha DE, Gaillard EA. What is the clinical significance of
filamentous fungi positive sputum cultures in patients with cystic
fibrosis? J Cyst Fibros 2013;12:187–193.
20. Coughlan CA, Chotirmall SH, Renwick J, Hassan T, Low TB, Bergsson
G, et al. The effect of Aspergillus fumigatus infection on vitamin D
receptor expression in cystic fibrosis. Am J Respir Crit Care Med
2012;186:999–1007.
21. Chaudhary N, Datta K, Askin FB, Staab JF, Marr KA. Cystic fibrosis
transmembrane conductance regulator regulates epithelial cell
response to Aspergillus and resultant pulmonary inflammation. Am J
Respir Crit Care Med 2012;185:301–310.
22. Amin R, Dupuis A, Aaron SD, Ratjen F. The effect of chronic infection
with Aspergillus fumigatus on lung function and hospitalization in
patients with cystic fibrosis. Chest 2010;137:171–176.
23. Fillaux J, Brémont F, Murris M, Cassaing S, Rittié JL, Tétu L, et al.
Assessment of Aspergillus sensitization or persistent carriage as a
factor in lung function impairment in cystic fibrosis patients. Scand J
Infect Dis 2012;44:842–847.
24. McMahon MA, Chotirmall SH, McCullagh B, Branagan P, McElvaney
NG, Logan PM. Radiological abnormalities associated with
Aspergillus colonization in a cystic fibrosis population. Eur J Radiol
2012;81:e197–e202.
25. Shoseyov D, Brownlee KG, Conway SP, Kerem E. Aspergillus
bronchitis in cystic fibrosis. Chest 2006;130:222–226.
26. de Vrankrijker AM, van der Ent CK, van Berkhout FT, Stellato RK,
Willems RJ, Bonten MJ, et al. Aspergillus fumigatus colonization in
cystic fibrosis: implications for lung function? Clin Microbiol Infect
2011;17:1381–1386.
27. Aaron SD, Vandemheen KL, Freitag A, Pedder L, Cameron W, Lavoie A,
et al. Treatment of Aspergillus fumigatus in patients with cystic
fibrosis: a randomized, placebo-controlled pilot study. PLoS One
2012;7:e36077.
28. Breuer O, Caudri D, Schultz A, Turkovic L, Ranganathan S, Stick S.
Aspergillus and early CF lung disease: guilty until proved innocent
[abstract]? Am J Respir Crit Care Med 2019;199:A4003.
696 American Journal of Respiratory and Critical Care Medicine Volume 201 Number 6 | March 15 2020
ORIGINAL ARTICLE
29. Sly PD, Brennan S, Gangell C, de Klerk N, Murray C, Mott L, et al.;
Australian Respiratory Early Surveillance Team for Cystic Fibrosis
(AREST-CF). Lung disease at diagnosis in infants with cystic fibrosis
detected by newborn screening. Am J Respir Crit Care Med 2009;
180:146–152.
30. D’Sylva P, Caudri D, Shaw N, Turkovic L, Douglas T, Bew J, et al.
Induced sputum to detect lung pathogens in young children with
cystic fibrosis. Pediatr Pulmonol 2017;52:182–189.
31. Stick SM, Brennan S, Murray C, Douglas T, von Ungern-Sternberg BS,
Garratt LW, et al.; Australian Respiratory Early Surveillance Team for
Cystic Fibrosis (AREST CF). Bronchiectasis in infants and preschool
children diagnosed with cystic fibrosis after newborn screening.
J Pediatr 2009;155:623–628, e1.
32. Rosenow T, Oudraad MC, Murray CP, Turkovic L, Kuo W, de Bruijne M,
et al.; Australian Respiratory Early Surveillance Team for Cystic
Fibrosis (AREST CF). PRAGMA-CF: a quantitative structural lung
disease computed tomography outcome in young children with
cystic fibrosis. Am J Respir Crit Care Med 2015;191:1158–1165.
33. Jubin V, Ranque S, Stremler Le Bel N, Sarles J, Dubus JC. Risk factors
for Aspergillus colonization and allergic bronchopulmonary
aspergillosis in children with cystic fibrosis. Pediatr Pulmonol 2010;
45:764–771.
34. Harun SN, Holford NHG, Grimwood K, Wainwright CE, Hennig S.;
Australasian Cystic Fibrosis Bronchoalveolar Lavage (ACFBAL)
Study Group. Pseudomonas aeruginosa eradication therapy and risk
of acquiring Aspergillus in young children with cystic fibrosis. Thorax
2019;74:740–748.
35. Tepper LA, Utens EM, Caudri D, Bos AC, Gonzalez-Graniel K,
Duivenvoorden HJ, et al. Impact of bronchiectasis and trapped air on
quality of life and exacerbations in cystic fibrosis. Eur Respir J 2013;
42:371–379.
36. Rosenow T, Ramsey K, Turkovic L, Murray CP, Mok LC, Hall GL, et al.;
AREST CF. Air trapping in early cystic fibrosis lung disease-does CT
tell the full story? Pediatr Pulmonol 2017;52:1150–1156.
37. Cowley AC, Thornton DJ, Denning DW, Horsley A. Aspergillosis and
the role of mucins in cystic fibrosis. Pediatr Pulmonol 2017;52:
548–555.
38. Sanders DB, Li Z, Brody AS, Farrell PM. Chest computed tomography
scores of severity are associated with future lung disease
progression in children with cystic fibrosis. Am J Respir Crit Care
Med 2011;184:816–821.
39. Turkovic L, Caudri D, Rosenow T, Ranganathan S, Hall GL, Stick S,
et al. Structural determinants of long term functional outcomes in
children with cystic fibrosis. Respirology 2017;22:55.
40. Bortoluzzi CF, Volpi S, D’Orazio C, Tiddens HA, Loeve M, Tridello G,
et al. Bronchiectases at early chest computed tomography in
children with cystic fibrosis are associated with increased risk of
subsequent pulmonary exacerbations and chronic pseudomonas
infection. J Cyst Fibros 2014;13:564–571.
41. Frost FJ, Nazareth DS, Charman SC, Winstanley C, Walshaw MJ.
Ivacaftor is associated with reduced lung infection by key cystic
fibrosis pathogens: a cohort study using national registry data. Ann
Am Thorac Soc 2019;16:1375–1382.
42. Hill AB. The environment and disease: association or causation? Proc
R Soc Med 1965;58:295–300.