Articles

Factors associated with failure of emergency department

management in children with acute moderate or severe
asthma: a prospective, multicentre, cohort study
Francine M Ducharme, Roger Zemek, Bhupendrasinh F Chauhan, Jocelyn Gravel, Dominic Chalut, Naveen Poonai, Marie-Claude Guertin,
Caroline Quach, Lucie Blondeau, Sophie Laberge, for the DOORWAY research group of the Pediatric Emergency Research in Canada (PERC) network

Summary
Background The management of paediatric asthma exacerbations is based on trials in children of all ages. Recent
studies from 2009 raised the possibility that preschoolers (younger than 6 years) with viral-induced wheezing and
children exposed to tobacco smoke might be at an increased risk of treatment failure. The study objective was to
identify factors associated with management failure in children presenting to the emergency department with
moderate or severe asthma exacerbations.
Methods We undertook a prospective, multicentre cohort study of children aged 117 years presenting to ve
emergency departments with moderate or severe asthma (dened as a Pediatric Respiratory Assessment Measure
[PRAM] of 4 to 12). Children received oral corticosteroids and severity-specic inhaled bronchodilator therapy. The
primary outcome was emergency department management failure (hospital admission, prolonged emergency
department therapy [8 h], or relapse within 72 h of discharge from the emergency department with admission to
hospital or prolonged emergency department stay). Viral cause was ascertained by PCR on nasopharyngeal specimens
and environmental tobacco smoke exposure by salivary cotinine concentration. This study is registered at ClinicalTrials.
gov (NCT02013076).
Findings Between Feb 14, 2011, and Dec 20, 2013, we screened 1893 children and enrolled 1012 eligible children. Of
those eligible children, 973 participants were included in the analysis. 165 (17%) of 965 children experienced
management failure in the emergency department, which was signicantly associated with viral detection (110 [19%]
of 579 participants with virus detection vs 46 [13%] of 354 participants without viral detection, odds ratio [OR] 157;
95% CI 104237), fever (24% vs 15%, 196; 132292), baseline PRAM (OR 138 per 1-point increase; 122156),
oxygen saturation of less than 92% (50% vs 12%, 394; 197789), and presence of symptoms between exacerbations
(21% vs 16%, 173; 113264). Age, salivary cotinine concentration, and oral corticosteroids dose were not
signicantly associated with management failure. Viral detection (67% vs 46%, p<00001) and fever (31% vs 16%,
p<00001) occurred more frequently in preschoolers than in older children. Viral detection was also associated with
reduced speed of recovery over the 10 days after discharge.
Interpretation In children presenting with moderate or severe asthma, viral detection, but not age, was associated
with failure of symptom management, independently from exacerbation severity (ie, baseline PRAM and oxygen
saturation), fever, and symptom chronicity (viral detection). Although it did not reach statistical signicance, the
association between treatment management failure and exposure to tobacco smoke warrants further investigation.
Funding Canadian Institutes of Health Research.

Introduction
Asthma is the most common chronic disease in childhood.1
The burden of asthma is signicantly higher in children
younger than 6 years, who account for more than 50% of
emergency department visits and are admitted to hospital
three times more often than older children.2 In patients
with moderate and severe exacerbations or poor response
to bronchodilator therapy, inhaled bronchodilators and
systemic corticosteroids are key components of acute
asthma management guidelines.35 However, not every
patient responds suciently well to avoid admission.
As inammation is believed to be the major component
of airway obstruction in patients with suboptimal
bronchodilator response, recent investigations have

focused on the response to oral corticosteroids. A landmark

systematic review of placebo-controlled trials6 conrmed a
25% reduction in admission rate in adult and paediatric
patients with moderate or severe acute asthma receiving
systemic corticosteroids in the emergency department with
a similar benet observed when including only paediatric
trials, in which approximately half of participants were
preschoolers (those younger than 6 years). In a subsequent
randomised trial of children aged 1060 months with mild
or moderate viral-induced wheezing,7 oral corticosteroids
were not superior to placebo for reducing length of hospital
stay, raising the possibility that young age and viral infection
might contribute to the null ndings. Critics of this trial
have suggested that inclusion of a majority of children

www.thelancet.com/respiratory Published online July 20, 2016 http://dx.doi.org/10.1016/S2213-2600(16)30160-6

Lancet Respir Med 2016

Published Online
July 20, 2016
http://dx.doi.org/10.1016/
S2213-2600(16)30160-6
See Online/Comment
http://dx.doi.org/10.1016/
S2213-2600(16)30202-8
Department of Pediatrics
(Prof F M Ducharme MD,
J Gravel MD, S Laberge MD),
Department of Social and
Preventive Medicine (Prof
F M Ducharme), and Montreal
Health Innovations
Coordinating Center
(M-C Guertin PhD,
L Blondeau MSc), University of
Montreal, Montreal, QC,
Canada; Research Centre,
Sainte-Justine University
Health Centre, Montreal, QC,
Canada (Prof F M Ducharme,
S Laberge); Department of
Pediatrics and Emergency
Medicine, University of Ottawa,
Ottawa, ON, Canada
(R Zemek MD); Childrens
Hospital of Eastern Ontario
Research Institute, Ottawa, ON,
Canada (R Zemek); Faculty of
Pharmacy and Centre for
Healthcare Innovation,
University of Manitoba,
Winnipeg, MB, Canada
(B F Chauhan PhD); Montreal
Childrens Hospital
(D Chalut MD, C Quach MD) and
Research Institute (C Quach),
McGill University Health Center,
Montreal, QC, Canada; and
Childrens Hospital, London
Health Sciences Center, London,
ON, Canada (N Poonai MD)
Correspondence to:
Dr Francine M Ducharme,
Department of Pediatrics,
Research Centre Sainte-Justine
University Health Centre,
3175 Cte Sainte-Catherine,
Room 7939, Montreal,
QC H3T 1C5, Canada
francine.m.ducharme@
umontreal.ca

Articles

Research in context
Evidence before this study
The acute management of preschoolers with asthma
exacerbations has been the subject of much debate with some
trials showing clear ecacy of standard therapy with
bronchodilators and oral corticosteroids and other trials
showing no such treatment eect. A potential reason for the
discordant ndings might be a varying distribution across trials
of factors associated with poor treatment response.
Added value of this study
We report the results of a large cohort study exploring the
determinants of emergency department management failure to
standard therapy in children aged 117 years with moderate or
severe asthma exacerbations. In addition to the expected acute

without asthma, low exacerbation severity, low oral

corticosteroid dose, and unclear justication for prolonged
hospital stay despite low severity, might have contributed to
the null ndings.8 Furthermore, although tobacco smoke is
associated with an increased number of respiratory
infections, increased severity of asthma exacerbations, and
reduced response to inhaled corticosteroids, the eect of
tobacco smoke on treatment response has not been
documented in children with asthma.9 The overrepresentation of boys in paediatric asthma trials raises the
possibility of a sex-related dierential response.6,7 None of
six aforementioned potential determinants of treatment
response, namely age, sex, viral trigger, severity, oral
corticosteroid dose, and tobacco exposure have been
formally explored in children diagnosed with acute asthma.
The objective of this study was to identify and quantify
factors associated with treatment failure with severityspecic bronchodilator therapy and oral corticosteroids
in children presenting to the emergency department with
moderate or severe asthma exacerbations.

Methods
Study design and participants
We undertook a multicentre, prospective cohort study of
children presenting with moderate or severe acute
asthma to one of ve participating Canadian paediatric
emergency departments. The institutional review boards
of the participating centres approved the study. Parents
provided informed consent for study participation and
assent was obtained in children who were able to
comprehend the nature of the study. The study was
conducted in accordance with Helsinki Good Clinical
Practice Guidelines.10
Individuals were eligible if they: were aged 117 years;
had a physician diagnosis of asthma, based on a previous
wheezing episode with signs of airow obstruction and
response to bronchodilators,11 three or more asthma-like
episodes (if <2 years old), or previous diagnostic lung
function tests, or a combination of these factors;12 presented
2

severity and chronicity of disease, we identied viral detection

and fever as important determinants of management failure,
whereas age, sex, and oral corticosteroid dose were not. Viral
detection was also associated with slower recovery over the
next 10 days.
Implications of all the available evidence
Viral respiratory infections might explain, in part, the apparent
poor response observed in some trials of preschoolers (younger
than age 6 years) presenting with viral-induced wheezing. Realtime viral testing might enable better identication of children
at higher risk of treatment failure and slower recovery, in whom
novel preventive or therapeutic approaches could be explored.

with a moderate or severe airway obstruction, dened as a

baseline Pediatric Respiratory Asthma Measure (PRAM)
score of 412 (a score of 0 is normal, 12 is the most severe);13
and were accompanied by a legal guardian with good
comprehension of English or French. Patients were
excluded if they presented with suspicion of bronchiolitis
or foreign body aspiration; another chronic respiratory
disorder; hypersensitivity to salbutamol, ipratropium
bromide, or prednisolone; or a contraindication to oral
corticosteroids. A central adjudicating committee reviewed
all participants who did not meet eligibility criteria, did not
receive or tolerate the target oral corticosteroid dose, or in
whom the emergency department physician was unsure of
the discharge diagnosis of asthma (<5 on a 7-point Likert
scale); two members of the study Central Adjudication
Committee independently reviewed each adjudicated case
and disagreements were resolved by consensus.

Treatment protocol
The study outline and schedule, which have been described
elsewhere,14 are depicted in gure 1. Participants were
assessed for symptom severity by a research nurse or
respiratory technician and assigned a PRAM score at
baseline. The baseline PRAM score enabled investigators
to classify patients as having a moderate (score of 47) or
severe (score of 812) exacerbation and to initiate the
severity-specic treatment. As per the evidence-based
paediatric guidelines,35 children received 20 mg/kg of oral
prednisone or prednisolone (maximum 50 mg)which,
if vomited, was replaced by 03 mg/kg of oral
dexamethasone15and severity-specic inhaled bronchodilator treatment with salbutamol, with or without
ipratropium bromide, according to the baseline PRAM.
Salbutamol was administered as 03 inhalation per kg of
100 g per inhalation (maximum ten inhalations) or
003 mL/kg of 05% salbutamol solution (maximum
1 mL). Ipratropium bromide was administered as three
inhalations of 20 g per inhalation or of 250 g nebules,
irrespective of the patients age. In the initial hour of

www.thelancet.com/respiratory Published online July 20, 2016 http://dx.doi.org/10.1016/S2213-2600(16)30160-6

Articles

treatment, children with a PRAM score of 47 received two

to three salbutamol inhalations every 2060 min and no
ipratropium bromide; those with a PRAM of 812 received
three salbutamol and ipratropium bromide inhalations at
20 min intervals. The only variations in the treatment
protocol allowed in the rst hour were limited to salbutamol
in children with a PRAM of 47 (one dose in one site and
two to three doses in four sites). After the initial hour,
salbutamol was administered once or twice hourly,
depending on clinical response. Co-interventions, such as
additional doses of ipratropium after the initial hour or
magnesium sulphate, were permitted and recorded. PRAM
was documented hourly until discharge or 4 h after the
administration of oral corticosteroids, whichever occurred
rst. Salivary samples and nasopharyngeal swabs were
taken after the initial hour of treatment.
As per its established protocol, one site was allowed by
the independent data monitoring committee to use
1 mg/kg of prednisone (015 mg/kg of dexamethasone).
Discharge drugs included a 4-day course of oral
prednisolone (1 mg/kg per day, maximum 50 mg) and
inhaled 2-agonists as needed; asthma controller drugs
(an inhaled corticosteroid with or without adjunct therapy or
a leucotriene receptor antagonist) were prescribed at the
physicians discretion.

Outcomes
The primary outcome, failure of emergency department
treatment management, was dened as hospital admission
for asthma, treatment in the emergency department lasting
8 h or more after administration of oral corticosteroids, or a
return to the emergency department within 72 h of
discharge leading to hospital admission or prolonged
emergency department stay of 8 h or longer. Secondary
measures of eectiveness in the emergency department
included: hospital admission for asthma; PRAM prole
(measured hourly until the decision to discharge home or
to hospitalise or 4 h after oral corticosteroids, whichever
occurred rst); time taken to reach a PRAM of 3; length of
active treatment (between rst and last salbutamol dose);
and proportion of patients with PRAM of 4 or more at 4 h
after oral corticosteroids. Secondary measures of resolution
of exacerbation, which were measured over the following
10 days, included: unscheduled emergency department
visits for asthma; cumulative symptom score, duration of
symptoms, cumulative number of salbutamol inhalations,
and duration of use of rescue 2-agonists use (measured
daily with the Asthma Flare-up Diary for Young Children);16
and parental functional status measured with the Eects of
a Young Childs Asthma Flare-up on Parents
Questionnaire.17

Severity-specic therapy
Optimal inhalations

Triage

Therapy
PRAM 47
Inhaled salbutamol
Oral corticosteroids

Delay from oral corticosteroid (h)

PRAM 812
Inhaled salbutamol
Inhaled ipratropium bromide
Oral corticosteroids
Measurements
PRAM
Salivary cotinine
Viral sampling

Figure 1: Study design and schedule of measurements

Baseline PRAM score classied patients into moderate (score of 47) or severe (score of 812) exacerbation groups
and serve to initiate the severity-specic therapy indicated by the red circles; optional inhalations are indicated by
light red circles. PRAM=Pediatric Respiratory Assessment Measure.

Asthma Quiz for Kidz20), tobacco exposure,21 asthma

controller use, fever, baseline oxygen saturation, season,
parents and physicians perceived triggers, and
emergency department management variation (minor
variations in the frequency of bronchodilators, dose, or
timing of oral corticosteroids; appendix p 2). Atopy refers
to any one of the following, as documented by the ISAAC
questionnaire: sneezing; runny or blocked nose in
absence of cold or hay fever; itching, rash, or eczema; or
rhinitis or hay fever.19 Childs symptom pattern reported
by parents as intermittent asthma if described as being
perfectly well between episodes with no cough, wheezing,
or diculty breathing ever, even when laughing, running,
or crying or as persistent asthma if described as having
intermittent or continuous symptoms between episodes,
that is, cough, wheezing or diculty breathing (even if
minimal for example when laughing, running, or crying).
Six key variables of interest were identied a priori
because of suspected or conrmed association with oral
corticosteroid response: age, sex, baseline severity of
exacerbation assessed on the 12-point PRAM,13 viral
detection, salivary cotinine concentration, and oral
corticosteroid dose. All children had a nasopharyngeal
aspirate or swab (Flocked swab, Copan Diagnostics, CA,
USA) inserted in 3 mL of viral transport media (UTM,
Copan Diagnostics, CA, USA) and frozen at 80C until
processing with a validated automated microarray
detection with PCR for 23 common respiratory viruses.22
A quantitative cotinine analysis with an enzyme
immunoassay kit (Salimetrics, PA, USA) was done on
saliva, which was kept frozen at 80C until processing.23
All laboratory measures were interpreted independently
of the knowledge of exposure and outcome.

See Online for appendix

Potential determinants of response

All potential determinants of emergency department
management failure were prespecied, comprising of
demographics, asthma phenotype,18 triggers and atopy,19
previous morbidity, asthma control (measured with the

Statistical analyses
The prevalence of the six key variables of interest and the
estimated risk of admission to hospital in the unexposed
(41%) variables lead to a sample size of 1200 to detect a

www.thelancet.com/respiratory Published online July 20, 2016 http://dx.doi.org/10.1016/S2213-2600(16)30160-6

Articles

clinically important risk dierence of 8% (odds ratios

[OR] of 14) in the emergency department management
failure for each factor (type I error=005) with a power of
80%.14 With a lower than expected proportion of failure in
51 (16%) of the initial 320 patients in September, 2012,
the sample size recalculated for a similar risk dierence
was 800 children. Estimating a 20% proportion of
dropouts or missing biological samples, we aimed to
recruit 1000 children.
For the primary outcome, we used bivariate and
multivariate stepwise logistic regression models to
examine the association with each potential variable. The
six key variables, namely, age, sex, baseline PRAM, viral
detection (yes or no), salivary cotinine (values of
<1 ng/mL, 1 ng/mL to <4 ng/mL, and 4 ng/mL),23 and
oral corticosteroids dose (mg/kg) were forced in the
original multivariate model. Other explored determinants
included host, episode, and management-related
variables (appendix p 2). The nal model resulted from
the stepwise selection with use of the six key variables

and other determinants, with investigational site added.

Missing cotinine samples were imputed as <1 ng/mL
when patients reported no exposure and 1 ng/mL to
<4 ng/mL when patients reported passive smoking. In a
sensitivity analysis, we repeated the analysis, including
only patients with measured salivary cotinine. Post hoc,
we explored multiple interactions and did a post-hoc
subgroup analysIs on age (<36 months vs 36 months)
and on viral cause (rhinovirus A, B, C, or rhinovirus or
enterovirus detected vs not detected). The factors
associated with the secondary outcomes were measured
at baseline (or before discharge, if pertaining to recovery
after admission or discharged home); these outcomes
were identied by a similar approach, with linear
regression models, logistic regression models, or Cox
proportional hazards models, according to the nature of
the outcome. All tests were two-sided with estimates
presented with 95% CIs. Analyses were done with SAS
software (version 9.3). No correction was done for multiple
testing. This study is registered at ClinicalTrials.gov
(NCT02013076).

Role of the funding source

1893 children were screened

696 were ineligible

53 no acute asthma
212 no suspected asthma
334 PRAM score 3
24 concomitant other chronic respiratory
disorder
16 suspicion of bronchiolitis or foreign body
aspiration
1 hypersensitivity to asthma drug
5 recently exposed to varicella
10 received live vaccination within 14 days
13 other contraindications
1 suspected or conrmed pregnancy
27 incomplete assessment or lost to follow-up

1197 were provisionally eligible

185 did not participate

1012 were enrolled

39 were excluded
3 withdrew during the emergency department
management
5 did not receive or tolerate oral corticosteroids
26 received an insucient or excessive dose of
oral corticosteroids
4 uncertainty about the diagnosis of asthma
1 ineligible because of baseline PRAM <4

973 were included in the analysis

Figure 2: Patient selection

The ow of patients is shown, from screening to analysis. PRAM=Pediatric
Respiratory Assessment Measure.

The funder of the study had no role in the study design,

data collection, data analysis, data interpretation, or
writing of the report. The corresponding author had full
access to all study data and had nal responsibility for
the decision to submit for publication.

Results
Between Feb 14, 2011, and Dec 20, 2013, we screened
1893 children. Of those screened, 696 (37%) children
were ineligible. Of the eligible children, 185 (15%) of
1197 declined to participate, leading to 1012 enrolled
children: three children withdrew and 36 were removed
because of serious protocol deviations (gure 2). Five
institutions recruited participants: Childrens Hospital
of Eastern Ontario (446 participants), Centre Hospitalier
Universitaire Sainte-Justine (299 participants), McGill
University Health Centre (184 participants), London
Health Sciences Centre (78 participants), and Centre
Hospitalier Universitaire de Laval (ve participants).
Baseline characteristics of non-participants were similar
to the 973 patients included in the analysis in age,
neighbourhood income (with postal code as a surrogate),
the number of courses of oral corticosteroids in the
preceding year, and baseline PRAM; slightly more girls
declined participation (93 [42%] of 224 non-participants
vs 328 [34%] of 972 participants, p=003). Most
participants were male preschoolers with intermittent
asthma who presented with a moderate exacerbation.
Few children had elevated salivary cotinine
concentrations and four reported smoking (table 1).
Although 83% of parents perceived a viral respiratory
infection as the most likely trigger of the exacerbation
(table 1), there was no signicant association between
perceived viral trigger and viral detection; 541 (94%) of

www.thelancet.com/respiratory Published online July 20, 2016 http://dx.doi.org/10.1016/S2213-2600(16)30160-6

Articles

Cohort (n=973)*
Demographic

Cohort (n=973)*
(Continued from previous column)

Age

Index exacerbation

15 years

730 (75%)

617 years

242 (25%)

Previous physician diagnosis of asthma

525 (54%)
799 (82%)

Male

645 (66%)

Previous episode with airow obstruction with

response to asthma medication

Female

328 (34%)

3 wheezing episodes if <2 years (n=372)

218 (59%)

Previous reversibility post-bronchodilator

documented by lung function

15 (2%)

Sex

Ethnicity
Caucasian

569 (58%)

Black

101 (10%)

Arabic

95 (10%)

Asian

72 (7%)

Mixed or other

136 (14%)

Risk factors
Age at rst diagnosis

15 (1025 years)

Hay fever, ever

135 (14%)

Eczema, ever

428 (44%)

Atopy

674 (69%)

Family history of asthma

531 (55%)

Environmental tobacco smoke

417 (43%)

In-utero smoking

105 (11%)

Morbidity before enrolment

Number of acute care visits in previous year

2 (03)

Patient with 1 course of oral corticosteroids in

previous year

473 (49%)

Patient with 1 hospital admission in previous year

152 (16%)

Asthma Quiz for Kidz score

3 (24)

Diagnostic criteria met (not mutually exclusive)

Positive provocation test

10 (1%)

Current diagnosis of asthma by emergency physician

903 (93%)

Fever documented at baseline

260 (27%)

Most important trigger (according to parents)

Viral infection

804 (83%)

Weather

87 (9%)

Allergen

38 (4%)

Smoke or irritants

14 (1%)

Others or unknown

28 (3%)

Most important trigger (according to physicians; n=964)

Viral infection

893 (93%)

Weather

20 (2%)

Allergen

21 (2%)

Smoke or irritants

8 (1%)

Others or unknown

22 (2%)

Asthma pathway initiated||

Moderate

647 (67%)

Severe

326 (35%)

Baseline PRAM score**

Asthma pattern
Pattern of asthma symptoms
Intermittent asthma

743 (76%)

Persistent asthma

229 (24%)

03

75 (8%)

47

705 (73%)

812

193 (20%)

Salivary cotinine (n=970)

Usual trigger
URTI only

704 (72%)

<1 ng/mL (no ETS exposure)

657 (68%)

Multiple or other triggers

244 (25%)

1 ng/mL to <4 ng/mL (undetermined exposure)

260 (27%)

4 ng/mL (meaningful exposure)

Preventive strategy before enrolment

Prescribed asthma controller
None

407 (42%)

Daily maintenance

282 (29%)

Episodic intake (when sick)

283 (29%)

Reported intake of asthma controller in previous month

None

690 (71%)

021 days

46 (5%)

2230 days

228 (23%)
(Table 1 continues in next column)

579 children with a viral detection compared with 319

(90%) of 354 of those without viral detection were
perceived by parents as having a viral infection (p=006).
579 (62%) of the 933 children with a valid sample had
one or more virus detected; rhinovirus, respiratory
syncytial virus, and metapneumovirus were the most
frequent (appendix p 4). Viral detection (471 [67%] of
701 vs 107 [46%] of 231, p<00001) and fever (221 [31%] of

53 (5%)

Data are n (%) or median (IQR). URTI=upper respiratory tract infection.

ETS=environmental tobacco smoke. *Unless otherwise stated. Asthma Quiz for
Kidz20 is a validated questionnaire measuring the number of indicators of poor
asthma control. A score of 0 is best, 6 is worst, and 2 is the validated cuto for
poor control. Inhaled corticosteroids as monotherapy (472 [84%] of
564 patients), in combination with antileukotrienes or long-acting 2-agonists
(80 [14%] of 564 patients), or antileukotriene as monotherapy (12 [2%] of
564 patients). Patients in whom the status for a given criterion was recorded as
unsure, not applicable, or not available were assumed not to have met the
criterion. With standard cutos for reversibility and for airway hyper-reactivity
on provocation testing, when the information was available in medical chart at
enrolment ||Severity-specic bronchodilator protocol was based on Pediatric
Respiratory Assessment Measure (PRAM) at enrolment.13 **Dened as latest
PRAM measured at or before oral corticosteroid administration. As patients had
received bronchodilators before oral corticosteroids administration, a small
proportion of children had improved suciently to have a PRAM of 3 or less at
that time. As a valid salivary sample for cotinine was not obtained in
130 (134%) of 973 children, we classied as no exposure those who reported no
household exposure, and as undetermined exposure those reporting some
household smoke exposure. With 579 (68%) of 843 patients with sample values
below 1 g/mL, the distribution of saliva data was too asymmetrical to allow for
any adequate transformation for continuous analysis.

Table 1: Baseline characteristics

www.thelancet.com/respiratory Published online July 20, 2016 http://dx.doi.org/10.1016/S2213-2600(16)30160-6

Articles

723 vs 39 [16%] of 240, p<00001) occurred signicantly

more frequently in preschoolers than in school-aged
children (those aged 6 years or older). Because we could
not ascertain relapse status in eight patients, the primary
outcome was documented in 965 children.
With regards to emergency department management,
most (92%) patients had perfect adherence to oral
corticosteroids (15% of target dose) and 789 (81%)
patients to the severity-specic bronchodilator protocol
(table 2). Overall, 598 (62%) of 973 patients received oral
corticosteroids within 60 min of triage, with a median
delay of 51 min (table 2). The PRAM score improved
greatly over time (appendix pp 56). At discharge,
881 (91%) of 973 children were prescribed oral
corticosteroids (median dose 1 mg/kg) for four additional
days; 678 (70%) of 973 children were prescribed a daily
controller medication, usually inhaled corticosteroids,
for 14 days or more (appendix p 7).
Failure of emergency department management
occurred in 165 (17%) of 965 patients (appendix p 9). Of
the six a-priori-specied variables of key interest, viral
detection and baseline PRAM were signicantly
associated with emergency department management
failure, whereas salivary cotinine concentration, oral
corticosteroid dose, sex, and age were not signicantly
associated (table 3). The association between salivary
cotinine concentration of 4 ng/mL or more and
management failure did not meet statistical signicance
(OR 199, p=00602; table 3). When the model was
repeated with age as a continuous outcome, age remained
not signicantly associated (OR 098; 95% CI 091106
Moderate
episodes (n=647)

Severe episodes
(n=326)

All episodes
(n=973)

635 (98%)

308 (95%)

943 (97)

12 (2%)

18 (6%)

Oral corticosteroids
Drug name
Prednisone/prednisolone
Dexamethasone

30 (3%)

Actual dose (mg/kg) in

prednisone equivalent*

192 (129200)

198 (187200)

195 (152200)

In site where intention to treat was 1 mg/kg

101 (099109)

100 (099102)

101 (099107)

In sites where intention to treat was

2 mg/kg

195 (181200)

199 (194200)

197 (186200)

Perfect adherence to prednisone and

dose 15%
Delay (min) between triage and oral
corticosteroids

595 (92%)
57 (3688)

299 (92%)
44 (3458)

894 (92%)
51 (3575)

Severity-specic bronchodilators
Salbutamol

559 (86%)

315 (97%)

874 (89%)

Ipratropium bromide

589 (91%)

287 (88%)

876 (90%)

Perfect adherence to bronchodilators

502 (78%)

287 (88%)

789 (81%)

Data are n (%) or median (IQR) unless otherwise stated.*Where 1 mg of prednisone=1 mg of prednisolone=015 mg of
dexamethasone. Although all doses were within acceptable range (within 20%) of target dose of 1 mg/kg or 2 mg/kg
depending on site, we dened perfect adherence as receiving prednisone or prednisolone, orally, at a dose within 15%
of the site-specic target dose (maximum dose of 50 mg).

Table 2: Adherence to corticosteroids and severity-specic inhaled bronchodilators in the rst hour
of therapy

per 1-year increase). In the nal multivariate model,

statistically signicant determinants of emergency
department management failure comprised viral
detection, baseline PRAM, oxygen saturation, fever, and
presence of symptoms between episodes. A correlation
was noted between low oxygen saturation and high
PRAM (p<00001), but the strength of association
between PRAM and management failure remained
similar after oxygen saturation was removed from the
model (OR 162 [95% CI 138189]).
The only observed interaction was between age and
fever, with fever being associated with an OR of 251
(95% CI 160394) in preschoolers, with no signicant
association in older children (appendix p 11). There was
no evidence of interaction between age and viral detection
(p=072). A subgroup analysis of children younger than
36 months (OR 153; 95% CI 080296) or 36 months
or older (OR 130; 95% CI 074227) showed a similar
eect size of viral detection, albeit with wider condence
intervals (appendix p 12).
Other variables were not signicantly associated with
emergency department management failure, including
viral-induced phenotype, parental and physicians
perception of viral trigger, adherence to therapy, or delay
between triage and oral corticosteroids (appendix p 13).
Post-hoc analysis showed no increased failure with
rhinovirus infection (OR 096; 95% CI 067139) or
with the subgroup of rhinovirus or enterovirus that
includes rhinovirus C and enterovirus D68 (100;
069144). However, viral detection was signicantly
associated with impaired response to oral corticosteroids
both in the emergency department and over the 10 days
after discharge (relapse to emergency department within
7 days was 35 [6%] of 562 with viral detection vs ten [3%]
of 342 with no viral detection; OR 220 [108451];
appendix p 16).

Discussion
In this large pragmatic cohort, only 17% of children with
moderate or severe asthma exacerbations experienced
management failure. Severity of exacerbation (baseline
PRAM and oxygen saturation), asthma symptoms
between episodes, viral detection, and fever were
strongly associated with emergency department
management failure, after adjustment for site. Because
of the large eect size that might be signicant,
meaningful exposure to tobacco smoke warrants further
evaluation. Age, sex, and oral corticosteroids dose were
not signicantly associated with the risk of management
failure.
Emergency department management failure was
signicantly associated with markers of baseline severity
and symptom chronicity. Symptom chronicity might be a
marker of persistent or uncontrolled asthma3 or might
signal a more dicult-to-treat phenotype.24 The observed
association between emergency department management
failure and baseline PRAM (OR 138 per 1-point

www.thelancet.com/respiratory Published online July 20, 2016 http://dx.doi.org/10.1016/S2213-2600(16)30160-6

Articles

increase),25 as well as with low oxygen saturation,26

conrms similar previous reports. Although associated
with oxygen saturation, the PRAM score provided
complementary information to oxygen saturation.
Strong independent risk factors of emergency
department management failure were viral detection and
fever, which increased the odds by more than 50% and
100%, respectively, suggesting that respiratory infection
of any cause is of importance. Nearly two-thirds of
patients, predominantly preschoolers, had one or more
viruses detected in their nasopharyngeal secretions,
mainly rhinovirus and respiratory syncytial virus.27 Viral
detection remained a strong factor of management
failure irrespective of age, whether tested as a continuous
or dichotomous variable, and without a signicant
interaction with other variables. Data from 2014 suggest
that a dierential viral cytotoxicity, combined with innate
immune deciency, might lead to increased viral
replication and cytokine expression, resulting in more
severe infection in susceptible individuals.28 The strong
association of fever with emergency department
management failure in preschoolers also points to other
unidentied respiratory pathogens, increased incidence
or virulence of certain microorganisms, or heightened
susceptibility of younger children, and deserves
replication. As age and sex were not associated with
management failure, an increased incidence of infection
or virus load in young children, rather than age, might
thus explain the poor oral corticosteroid response
observed in some trials of preschoolers with viralinduced wheezing.7,29 Indeed, our ndings provide
evidence indicating that respiratory infections might be
associated with some level of resistance to asthma
therapy in the emergency department, with slower
resolution of symptoms, and more relapses after
discharge.
Of particular interest was whether children exposed to
tobacco smoke have a blunted oral corticosteroid
response, as reported in adult smokers.30 A salivary
cotinine concentration of 4 ng/mL or more, conrmed in
5% of participants, was associated with a two times
increase in emergency department management failure,
although the association did not reach statistical
signicance. However, this study was insuciently
powered to rmly exclude a deleterious eect of a
meaningful exposure to tobacco smoke in view of the low
observed prevalence of meaningful exposure; this
relationship deserves further evaluation in a population
with higher exposure.
The ceiling eect of the maximum dose of prednisone
and the inclusion of one site administering 1 mg/kg
prednisone oered an opportunity to examine the eect
of oral corticosteroids dose on response, which was not
signicantly associated with emergency department
management failure. With a post-hoc power of 80% to
detect an OR of 13, the negligible eect size
(OR 102; 95% CI 083125) with a narrow condence

Participants Failure of
(n=973)
emergency
department
management
(n=965; %)

Key
determinants*
adjusted odds
ratio (95% CI;
n=912)

Final models
adjusted odds
ratio (95% CI;
n=916)

Final model
p value

Age*
5 years

730 (75%)

120 (17%)

6 years

242 (25%)

45 (19%)

Male

645 (66%)

109 (17%)

Female

328 (34%)

56 (17%)

127 (077210)

Sex*

094 (063141)

Symptoms between exacerbations

No symptoms

743 (76%)

117 (16%)

Symptoms

229 (24%)

47 (21%)

159 (102247)

173 (113264)

208 (139313)

196 (132292)

0011

Fever
Yes

260 (27%)

63 (24%)

No

704 (72%)

102 (15%)

00009

PRAM at baseline*
Per 1-point
increase

138 (122156)

138 (122156)

<00001

O2 saturation at baseline
95%

714 (73%)

88 (12%)

9294%

200 (21%)

48 (24%)

151 (094242)

150 (094238)

00887

59 (6%)

29 (50%)

379 (189761)

394 (197789)

00001

<92%

Viral trigger*||
Not detected

354 (36%)

46 (13%)

Detected

579 (60%)

110 (19%)

161 (106245)

157 (104237)

00312

Cotinine*,**
657 (68%)

101 (15%)

1 ng/mL <4 ng/mL 260 (27%)

<1 ng/mL

48 (19%)

104 (067161)

4 ng/mL

15 (28%)

199 (097408)

093 (064137)

53 (5%)

Oral corticosteroids dose*

Per 1 mg/kg
increase

Sites
1

286 (29%)

59 (21%)

173 (18%)

16 (9%)

037 (018077)

041 (021081)

00096

79 (8%)

18 (23%)

088 (044176)

092 (046184)

08208

435 (45%)

72 (17%)

057 (036091)

059 (038094)

00250

3 and 5 combined
4

Data are n (%) unless otherwise stated. PRAM=Pediatric Respiratory Assessment Measure. *The six a-priori variables of
key interest were added to the model on the basis of a previous hypothesis. Final model adjusted for symptoms
between exacerbations, fever, baseline PRAM, oxygen saturation, viral trigger, and sites. An additional analysis was
done by repeating the nal model with key determinants, with age as a continuous variable. The adjusted OR (95% CI)
for age was 127 (077210) per 1 year of age increase. Documented body temperature of more than 383C rectal or
equivalent. PRAM is a 12-point validated score (0best to 12worst).13 ||Viral detection by PCR on a nasopharyngeal
swab or aspirate.22 **Documented by enzyme immunoassay (Salimetrics, PA, USA) on saliva.23 For unavailable
specimen, the information was imputed as follows: no reporting household smoking was categorised as cotinine less
than 1 ng/mL and reported household smoking as 1 ng/mL to less than 4 ng/mg.23 Reported in prednisone or
prednisolone in mg/kg equivalent based on a conversion factor of 066 of oral dexamethasone to prednisone. Site 2
used 1 mg/kg of prednisone.

Table 3: Analysis of determinants of failure of emergency department treatment

interval would suggest a at doseresponse between

1 mg/kg and 2 mg/kg of prednisolone. The overall low
emergency department management failure rate
supports the benet of guidelines-recommended therapy,
including oral corticosteroids, in children with moderate

www.thelancet.com/respiratory Published online July 20, 2016 http://dx.doi.org/10.1016/S2213-2600(16)30160-6

Articles

and severe exacerbations, even with a predominance of

preschoolers.3
The study must be interpreted with respect to the
following limitations. In view of international guidelines
for the management of moderate and severe asthma
exacerbations, we felt it would be unethical to conduct a
placebo-controlled trial of oral corticosteroids in this
high-morbidity population.35 As the major threat to
validity of a cohort study is confounding-by-indication
that is, change in therapy according to perceived needs
all patients were treated according to a standardised
severity-specic protocol; the few patients with severe
protocol deviations were removed by the adjudication
committee. Without a control group, we cannot rmly
conclude as to the ecacy of management or whether
the observed treatment resistance was to oral
corticosteroids, inhaled bronchodilators, or both.
Although we cannot rule out the possibility that
unmeasured dierences in management might have
been important, variations in the frequency of
bronchodilators, dose, or timing of oral corticosteroids
were not retained in the nal model. Adherence to oral
corticosteroids after discharge was not assessed but was
unlikely to have a major eect on emergency department
management failure because of the small proportion
(27%) of return visits. Although less likely in children
aged 1 year and older,7 fever could have been a reason for,
rather than a determinant of, admission; however, viral
detection was unknown at disposition. We did not
analyse specimens for bacterial infection. Atopy was
ascertained by questionnaire; few children had had
previous skin prick tests, and we did not measure
specic serum IgE. Although there is no universally
accepted denition of asthma in young children, we
adopted the Canadian Thoracic and Paediatric Societies
diagnostic criteria for preschoolers;11 we enrolled
children with at least one previous episode of airow
obstruction and reversibility in response to asthma
medication with no suspicion of another respiratory
disorder, who were aged 1 year or older (to reasonably
exclude bronchiolitis). All participants presented with
signicant airow obstruction (PRAM 4) and most
showed a signicant improvement in PRAM after
treatment in the emergency department, conrming
reversibility. Repeating the analysis on children younger
than 36 months did not change the strength of the
association, which supports the robustness of ndings
in the younger children in whom a diagnosis of asthma
might be confused with bronchiolitis or lower respiratory
infection. Few screened patients (40 [2%] of 1893) were
excluded because of other respiratory conditions and
only four patients were withdrawn by the adjudicating
committee because of diagnostic uncertainty. Although
the ndings improve our understanding of the
determinants of emergency department management
failure, the study was not designed to oer a clinical
predictive model. Finally, the pragmatic recruitment
8

resulted in a heterogeneous population, which provided

a reference group for each determinant and increased
generalisability.
In children with moderate or severe acute asthma
treated with severity-specic bronchodilators and oral
corticosteroids, the risk of emergency department
management failure is low. Viral detection and fever
were associated with increased management failure,
independently from exacerbation severity and symptom
chronicity: of interest, management failure was not
signicantly associated with age, sex, and oral
corticosteroids dose. The eect of meaningful exposure
to tobacco smoke deserves further evaluation.
Respiratory infection rates, rather than age, might thus
explain the poor response observed in some previous
trials of preschoolers with viral-induced wheezing.7
Contributors
FMD conceptualised the research design, wrote the research protocol,
secured the funding, and coordinated the project team. M-CG was
responsible for the sample size calculation and statistical methods and,
with LB, directed the statistical analyses. FMD had access to all the data
in the study and takes responsibility for the integrity of the data and
M-CG assumes responsibility for the data analysis. BFC contributed as a
postdoctoral fellow in the conduct of the study. SL coordinated the
Central Adjudicating Committee. As site investigators, RZ and FMD,
assisted by JG, DC, and NP, coordinated the study at their site,
supervised patient recruitment, and monitored protocol adherence.
RZ, JG, DC, NP, SL, and CQ provided feedback on the protocol,
interpretation of the study results, and on the manuscript. All authors
read and approved the nal version.
Declaration of interests
FMD received unrestricted donations from Boehringer Ingelheim, Merck
Canada, GlaxoSmithKline, and Novartis, a research grant from Merck,
and is serving on an advisory board of Boehringer Ingelheim, outside of
this Article. CQ received grants from Pzer, Sage, AbbVie, and GSK,
outside of this Article. All other authors declare no competing interests.
Acknowledgments
The study was funded by the Canadian Institutes of Health Research
(#102547) under the title Determinants Of Oral corticosteroid
Responsiveness in Wheezing Asthmatic Youth (DOORWAY). The work
was presented in part at the American Thoracic Societys meeting,
Denver, USA, in May, 2015, the Pediatric Academic Societies Conference,
San Diego, USA, in May, 2015, the European Academy of Allergy and
Clinical Immunology meeting, Barcelona, Spain, in June, 2015 and the
Canadian Pediatric Society, Toronto, Canada, in June, 2015. We
acknowledge the support of the Fonds de la Recherche du Qubec en
Sant for the infrastructure support provided to the Research Institutes of
the Centre Hospitalier Universitaire Sainte-Justine (CHUSJ), the McGill
University Health Centre (MUHC), and the Centre Hospitalier
Universitaire de Laval (CHUL). We thank parents of children enrolled in
this study; Chantal Guimont, site investigator at the CHUL as well as
research coordinators, specically Veronique Brisson, Marie-France Goyer
(CHUSJ), Diane Laforte, Dorothy McKelvey (MUHC), Holly Schnare and
Candice McGahern (Childrens Hospital of Eastern Ontario), and
Cindy Langford (London Health Sciences Centre) and
Annie-Claude DAnjou (CHUL); Lynda Robitaille and Jacques Corbeil of
the Centre de Gnomique de Qubec, Research Centre (CHUL) for
conducting and supervising the PCR analysis on nasopharyngeal samples;
Anita Franco and Dr Edgard Delvin, for conducting and interpreting,
respectively, the salivary cotinine assay; the members of the Central
Adjudicating Committee, namely Maala Bhatt, Martha McKinney, and
Sophie Laberge; the members of the Data Monitoring Committee, namely
Benoit Msse, David Johnson, and Ian MacLusky; Annie Thort for
assistance in manuscript preparation; and Suzanne Schuh and
Jacques Lacroix for their editorial comments.

www.thelancet.com/respiratory Published online July 20, 2016 http://dx.doi.org/10.1016/S2213-2600(16)30160-6

Articles

References
1
Public Health Agency of Canada. Life and breath: respiratory disease
in Canada. Ottawa: Public Health Agency of Canada, 2007.
2
Lougheed MD, Garvey N, Chapman KR, et al. The Ontario Asthma
Regional Variation Study: emergency department visit rates and the
relation to hospitalization rates. Chest 2006; 129: 90917.
3
Global Initiative for Asthma Prevention. Global strategy for asthma
management and prevention. Global Initiative for Asthma. 2016.
http://ginasthma.org/2016-gina-report-global-strategy-for-asthmamanagement-and-prevention (accessed July 7, 2016).
4
British Thoracic Society Scottish Intercollegiate Guidelines
Network. British guideline on the management of asthma.
Thorax 2014; 69 (suppl 1): i1192.
5
Ortiz-Alvarez O, Mikrogianakis A, Canadian Paediatric Society,
Acute Care Committee. Managing the paediatric patient with an
acute asthma exacerbation. Paediatr Child Health 2012; 17: 25156.
6
Rowe BH, Spooner C, Ducharme FM, Bretzla JA, Bota GW.
Early emergency department treatment of acute asthma with
systemic corticosteroids. Cochrane Database Syst Rev 2008;
4: CD002178.
7
Panickar J, Lakhanpaul M, Lambert PC, et al. Oral prednisolone for
preschool children with acute virus-induced wheezing. N Engl J Med
2009; 60: 32938.
8
Ducharme FM, Zemek RL, Schuh S. Oral corticosteroids in
children with wheezing. N Engl J Med 2009; 360: 1674.
9
Chilmonczyk BA, Salmun LM, Megathlin KN, et al. Association
between exposure to environmental tobacco smoke and
exacerbations of asthma in children N Engl J Med 1993; 328: 166569.
10 World Medical Association. World Medical Association Declaration
of Helsinki: ethical principles for medical research involving
human subjects. JAMA 2013; 310: 219194.
11 Ducharme F, Dell S, Radhakrishnan D, et al. Diagnosis and
management of asthma in preschoolers: a Canadian Thoracic
Society and Canadian Paediatric Society position paper.
Can Respir J 2015; 22: 13543.
12 American Thoracic Society. Standardization of spirometry.
Am J Respir Crit Care Med 1995; 152: 110736.
13 Ducharme FM, Chalut D, Plotnick L, et al. The pediatric respiratory
assessment measure: a valid clinical score for assessing acute
asthma severity from toddlers to teenagers. J Pediatr 2008;
152: 47680.
14 Ducharme FM, Zemek R, Gravel J, et al. Determinants of oral
corticosteroid responsiveness in wheezing asthmatic youth
(DOORWAY): protocol for a prospective multicentre cohort study of
children with acute moderate-to-severe asthma exacerbations.
BMJ Open 2014; 4: e004699.
15 Keeney GE, Gray MP, Morrison AK, et al. Dexamethasone for acute
asthma exacerbations in children: a meta-analysis. Pediatrics 2014;
133: 49399.
16 Ducharme FM, Jensen ME, Mendelson MJ, et al. Asthma are-up
diary for young children to monitor the severity of exacerbations.
J Allergy Clin Immunol 2015; 137: 74449.

17

18

19

20

21

22

23

24

25

26

27

28

29

30

Jensen ME, Mendelson MJ, Desplats E, Zhang X, Platt R,

Ducharme FM. Caregivers functional status during a young childs
asthma exacerbation: a validated instrument. J Allergy Clin Immunol
2015; 137: 78288.
Bacharier LB, Boner A, Carlsen KH, et al. Diagnosis and treatment
of asthma in childhood: a PRACTALL consensus report. Allergy
2008; 63: 534.
Asher MI, Keil U, Anderson HR, et al. International study of
asthma and allergies in childhood (ISAAC): rationale and methods.
Eur Respir J 1995; 8: 48391.
Ducharme FM, Davis GM, Noya F, Rich H, Ernst P. The Asthma
Quiz for Kidz: a validated tool to appreciate the level of asthma
control in children. Can Respir J 2004; 11: 54146.
OLoughlin J, Tarasuk J, Difranza J, Paradis G. Reliability of selected
measures of nicotine dependence among adolescents.
Ann Epidemiol 2002; 12: 35362.
Raymond F, Carbonneau J, Boucher N, et al. Comparison of
automated microarray detection with real-time PCR assays for
detection of respiratory viruses in specimens obtained from
children. J Clin Microbiol 2009; 47: 74350.
Salimetrics. Guidelines for interpreting cotinine levels:
United States. https://www.salimetrics.com/assets/documents/
Spit_Tips_-_Cotinine_Guidelines.pdf (accessed Oct 15, 2015).
Dougherty RH, Fahy JV. Acute exacerbations of asthma:
epidemiology, biology and the Exacerbation-prone phenotype.
Clin Exp Allergy 2009; 39: 193202.
Bhogal SK, McGillivray D, Bourbeau J, Benedetti A, Bartlett S,
Ducharme FM. Early administration of systemic corticosteroids
reduces hospital admission rates in children with moderate and
severe asthma exacerbation. Ann Emerg Med 2012; 60: 8491.
Mehta SV, Parkin PC, Stephens D, Keogh KA, Schuh S.
Oxygen saturation as a predictor of prolonged, frequent
bronchodilator therapy in children with acute asthma. J Pediatr 2004;
145: 64145.
Tsukagoshi H, Ishioka T, Noda M, Kozawa K, Kimura H.
Molecular epidemiology of respiratory viruses in virus-induced
asthma. Front Microbiol 2013; 4: 278.
Nakagome K, Bochkov YA, Ashraf S, et al. Eects of rhinovirus
species on viral replication and cytokine production.
J Allergy Clin Immunol 2014; 134: 33241.
Beigelman A, King TS, Mauger D, et al. Do oral corticosteroids
reduce the severity of acute lower respiratory tract illnesses in
preschool children with recurrent wheezing? J Allergy Clin Immunol
2013; 131: 151825.
Chaudhuri R, Livingston E, McMahon AD, Thomson L, Borland W,
Thomson NC. Cigarette smoking impairs the therapeutic response
to oral corticosteroids in chronic asthma. Am J Respir Crit Care Med
2003; 168: 130811.

www.thelancet.com/respiratory Published online July 20, 2016 http://dx.doi.org/10.1016/S2213-2600(16)30160-6