RANDOMIZED, DOUBLE-BLIND, PLACEBO-CONTROLLED TRIAL OF ORAL
ALBUTEROL IN INFANTS WITH MILD-TO-MODERATE ACUTE
HEMA PATEL, MD, MSC, SERGE GOUIN, MDCM, AND ROBERT W. PLATT, MSC, MS, PHD
To determine whether oral albuterol is effective in reducing symptomatology of acute viral bronchiolitis
in infants with mild-to-moderate illness.
In a randomized, double-blind trial, previously well infants were randomized upon discharge from the emergency
department to receive either albuterol (0.1 mg/kg/dose) three times per day or placebo three times per day for 7 days. Daily stan-
dardized telephone interviews were conducted for as long as 14 days. The primary outcome was the time to resolution of illness.
Secondary outcomes included time to normal feeding, normal sleeping, quiet breathing, resolved cough, and coryza.
We studied 129 infants (albuterol, n = 64; placebo, n = 65). Baseline characteristics were similar between groups. The
overall mean age was 5.3 months, 60% were male, and 49 of 61 tested infants were positive for respiratory syncytial virus. The
median (95% confidence interval) time to resolution of illness (days) was similar: albuterol, 9.0 (8–13); placebo, 8.0 (7–9); P
.3) (log-rank test). There were no significant group differences in any secondary outcome. Health care revisit and admission
rates were similar between groups.
No significant group differences in either primary or secondary outcomes in infants treated with oral albuterol
versus placebo were found. The widespread use of oral albuterol in this patient group is not recommended. (J Pediatr
lthough acute viral bronchiolitis is the most common lower respiratory tract
condition in infancy,1–3 very little is known about effective therapy for infants withmild-to-moderate respiratory distress, the vast majority of affected infants.
Standard pediatric texts recommend oral bronchodilators in symptomatic infants who donot require hospital admission.4–6 Surveys and pharmacy prescription record databases showthat albuterol (Ventolin, GlaxoSmithKline, Mississauga, Ontario, Canada) is commonlyprescribed to infants with mild-to-moderate bronchiolitis, usually for 5 to 7 days.7
However, there is little evidence to support this widespread practice. To our knowl-
the Depar tment of Epidemiology andBiostatistics, McGill University, and the
edge, only two trials evaluating the use of oral bronchodilators in bronchiolitis have been
published.8,9 Both studies examined the short-term effects (at 30 and 60 minutes) of one
dose of oral albuterol and found no difference relative to placebo-treated infants in oxygen
Supported by the Canadian Association
saturation, respiratory rate, or clinical score.8,9 Given the estimated hundreds of thousands
of infants receiving oral albuterol in North America each bronchiolitis season, we per-
treal Children’s Hospital Research Insti-tute.
formed a randomized controlled trial to assess the effectiveness of this intervention.
Thus, the objective of the study was to determine whether oral albuterol was effec-
clinicien with the Fonds de la rechercheen santé du Québec (FRSQ).
tive in reducing the time to resolution of illness (ROI) and specific symptom severity of
outpatient infants with mild-to-moderate acute viral bronchiolitis. We hypothesized that
infants treated with oral albuterol would have a shorter duration of illness.
Reprint requests: Hema Patel, MD, MSc,FRCPC, Montreal Children’s Hospital,2300 Tupper St, Room A216, Montreal,Quebec, Canada H3H 1P3. E-mail:[email protected]
Respiratory Distress Assessment Instrument
dosing units and were numerically coded for the use of thestudy recruitment team. All study personnel and participants
were blinded to treatment assignment for the duration of the
This study was a double-blind, randomized, placebo-
study. To evaluate blinding, parents and the research nurse
controlled, parallel group trial. It was approved by the Scien-
were questioned on the last study day about which therapy
tific Review Committee of the Montreal Children’s Hospital
Research Institute and by the Montreal Children’s HospitalEthics Review Board.
The study medication was administered by using a nee-
dle-less syringe that had been labeled at the calculated volume
Infants were recruited during three successive bronchi-
by the study recruitment team. The first dose of medication
olitis seasons (November 1998–March 2001) from the emer-
was given in the ED before discharge, and the method of ad-
gency department (ED) at our university-affiliated pediatric
ministration was taught to the caregiver.
Daily structured telephone interviews with the same
caregiver were conducted until the infant had complete symp-
tom resolution or for a duration of 14 days, whichever event
Eligible infants included all previously well infants (≤12
was earlier. All infants had a final follow-up call 14 days after
months old) with a clinical diagnosis of acute viral bronchioli-
entering the study. The standardized bilingual interview was
tis who were discharged directly home after an ED visit dur-
pretested for face validity, clarity, and comprehensiveness. Do-
ing the study period. Bronchiolitis
was clinically defined as the
mains of interest included (1) overall health status of the in-
first episode of wheezing in an infant and evidence of an acute
fant, (2) adherence with the prescribed therapy, (3) presence
respiratory tract infection (coryza or body temperature >38°C
and severity of bronchiolitis symptoms, (4) feeding and sleep-
rectal or cough). Infants were considered for discharge direct-
ing pattern of the infant, (5) return healthcare visits and hos-
ly home from the ED if they had (1) adequate hydration, (2)
pital admissions, (6) parental life disruptions, and (7) adverse
no lethargy, (3) hemoglobin oxygen saturation ≥95% in room
air, and (4) mild respiratory distress (respiratory rate ≤70breaths/min, heart rate ≤200 beats/min, and mild accessory
Measurements and Outcomes
muscle use). A working telephone and a consistent primary
Baseline measures included demographic features, age,
caregiver were required for all participating infants.
sex, weight, body temperature, history of atopy in the infant
Infants were excluded from the trial if they were older
or first degree relatives, passive cigarette smoke exposure, du-
than 12 months, had a previous history of wheezing or home
ration and symptoms of illness, viral etiology, vital signs, he-
bronchodilator use, or required direct hospital admission from
moglobin oxygen saturation in quiet rest, and degree of
the ED. Infants were also excluded if they had (1) a gestation-
respiratory distress. Respiratory distress was measured by
al age at birth of <34 weeks, (2) an underlying chronic cardiac
using the Respiratory Distress Assessment Instrument
or chronic pulmonary disease (eg, bronchopulmonary dyspla-
(RDAI), a 17-point categorical score developed by Lowell et
sia), (3) immunocompromise, (4) a history of immunoprophy-
al.10 This score, which measured wheezing and chest wall re-
laxis therapy (ie, respiratory syncytial virus immunoglobulin
tractions, was chosen because of its face validity, high inter-
or RSV monoclonal antibody therapy), or (5) a parent not flu-
rater and intrarater reliability, and discriminative ability.10–15
The primary outcome of interest was the time to ROI,
the period from study enrollment to the time that the infant
Randomization, Allocation, and Treatment Groups
returned to baseline health status, as determined by the prin-
Computer-generated randomization (SAS version 8.0,
cipal caregiver. This type of outcome measure has been used
using PROC PLAN, SAS Institute, Cary, NC) within blocks
in studies of outpatients with viral croup, another common
of six subjects was used to assign patients to one of two treat-
pediatric respiratory tract infection.16,17 The ROI was select-
ment strategies: (1) oral albuterol (0.1 mg/kg/dose) or (2) oral
ed as a practical, common-sense reflection of effectiveness of
placebo. Infants were treated three times daily for a maximum
the intervention. The ROI was scored by using the 4-point
of 7 days or until complete resolution of bronchiolitis symp-
ordinal scale developed by Cruz et al.16 On this scale, parents
toms, whichever event was earlier. Medications were supplied
rated the infants’ symptoms globally in one of the following
to caregivers on discharge from the ED. Treatment was allo-
categories: (1) symptoms worse, (2) symptoms same, (3)
cated by the department of pharmacy, with the code for med-
symptoms improved, or (4) symptoms resolved.16 The time to
ication allocation held by our study pharmacist, who had no
symptom resolution—that is, a score of 4—was compared be-
Secondary outcomes of interest included duration of
cough, coryza, noisy breathing, time to normal feeding, and
Both study syrups were identically clear, colorless, and
time to normal sleeping. Normal feeding
and normal sleeping
citrus flavored. Equal volumes of medication, dispensed in
were defined as each infant’s pre-illness pattern. Duration of
opaque brown bottles, were prepared in advance with standard
was defined as the time from study enrollment to the
time when the infant coughed “rarely” or “infrequently” versus
(Fig 1). Of the 231 potentially eligible infants, 60 parents de-
“often” or “always.” The revisit and the hospital admission
clined participation in the study, and another 42 infants were
rates were compared between groups. We monitored all pa-
not approached for study at the attending ED physician’s re-
tients for adverse events and systematically documented
quest. Age and sex distribution, baseline respiratory rate, and
episodes of tremor and vomiting but not tachycardia, because
nasopharyngeal aspirate results of these excluded 102 infants
it could not be reliably measured by parents at home. Changes
were similar to those of the study group. Of the remaining 129
in the infants’ condition and potential adverse events were
infants, 64 were allocated to receive albuterol, and 65 received
monitored by using a 24-hour on-call service for parents.
Infants in each treatment group were similar in baseline
Sample Size and Statistical Methods
characteristics and illness severity at ED presentation (Table).
The primary analysis was performed according to inten-
The mean age in months (SD) was 5.4 (3.1) for albuterol and
tion-to-treat guidelines for enrolled patients, regardless of
5.1 (2.6) for placebo. Nasopharyngeal aspirate (NPA) results
withdrawals or protocol deviations. The Student t
were available from 61 infants; RSV was the pathogen found
used to compare the mean time to ROI (in days) between
most frequently (albuterol, 26/32, 81%; placebo, 23/29; 79%).
groups. To detect a mean difference of 2 days in ROI, with an
Parainfluenza, influenza, and adenovirus strains were found in
α = .05 and power = 80%, 126 infants were required (63 in
the remainder of positive NPA results.
each group). This calculation was based on a mean duration of
The mean (SD) time to ROI in days was similar: al-
illness of 8 days (range, 7–10 days), with an estimated stan-
buterol, 8.9 (4.0), and placebo, 8.4 (3.7) (t
= .5); because
dard deviation of 4 days. Kaplan-Meier survival log-rank tests
of the skewed distribution, the median (95% CI) time in days
were conducted to analyze the difference in time to ROI be-
was considered a more appropriate presentation of results: al-
tween groups. Similar techniques were applied to all second-
buterol, 9.0 (8–13), and placebo, 8.0 (7–9) (log-rank test P
ary outcomes. Cox proportional hazards regression was used
.30) (Fig 2, A
). There were no significant group differences in
to evaluate potential confounders.18 Group differences in the
the secondary outcomes: duration of poor feeding, cough,
baseline characteristics were evaluated descriptively, and clini-
coryza, and noisy breathing (Fig 2, B
). Similarly, there were
cally relevant group differences were examined for potential
no significant differences in the median time in days (95% CI)
confounding. Interobserver reliability for the RDAI was test-
normal infant sleeping (albuterol, 3 [2–4], vs placebo, 3 [3–4];
ed by using the weighted κ coefficient; a value of κ ≥0.8 was
log-rank test P
= 0.15) or normal parental sleeping (albuterol,
3 [2–4], vs placebo, 3 [3–6]; log-rank test P
There were two withdrawals in each treatment group.
In the albuterol group, one infant was reassessed on day 2 of
During the study period, 1039 infants with acute viral
follow-up, and the study medication was stopped at the dis-
bronchiolitis were evaluated in the ED and were discharged
cretion of a nonstudy physician. This child was prescribed oral
directly home. Of these, 808 were not eligible for study
albuterol syrup for 1 week, and follow-up telephone assess-
Randomized, Double-Blind, Placebo-Controlled Trial of Oral Albuterolin Infants with Mild-to-Moderate Acute Viral Bronchiolitis
Table. Baseline characteristics
Albuterol (n = 64)
Placebo (n = 65)
Median days ill before ED presentation, n (quartiles)
*Percent hemoglobin oxygen saturation.HR
, Heart rate; RR
, respiratory rate.
ments were discontinued at the parent’s request. A second
fore 1 week, two because of recurrent vomiting, one because of
child in the albuterol group was withdrawn when the parents
shakiness and tremor, and two for unspecified reasons. While
reported on day 2 that the infant had previously received oral
treatment assignment was still blinded, three infants in the al-
albuterol. In the placebo group, one patient was reassessed on
buterol group were started on albuterol by metered dose in-
day 3 by a nonstudy physician and the study medication dis-
haler within the study period. In the placebo group, one infant
continued. This infant was started on inhaled albuterol and
discontinued the study medication on day 3 for unspecified
fluticasone, and the family declined follow-up telephone as-
reasons. A second infant in the placebo group was started on
sessments. The second child withdrawn from the placebo
inhaled albuterol during the study period.
group received one dose of the study medication in the ED;
Secondary analyses excluding infants withdrawn, in-
the family decided immediately after that they did not want to
fants admitted to the hospital, and infants with major proto-
col deviations, had results similar to the intent-to-treat results.
There were four infants assigned to albuterol who even-
No significant group differences were found in any compari-
tually required admission to the hospital for respiratory dis-
son of baseline characteristics or outcome.
tress, two on day 2 and two on day 3 of follow-up. There were
The median number (quartile range) of health care re-
five infants in the placebo group who required hospital admis-
visits within 14 days of study enrollment was similar between
sion, two on day 2 and one infant each on days 3, 4, and 8.
groups: albuterol, 1.0 (0–1), and placebo, 0 (0–2). Fifteen pri-
Study medications were discontinued at the time of admis-
mary caregivers in each treatment group reported missing work
sion, but follow-up measures were collected in all nine infants.
for >1 day during the study period (range, (albuterol) 1–9 days,
All study protocol deviations were also recorded. A
(placebo) 1–8 days). Trembling and vomiting were potential
major protocol deviation
was defined as one in which the infant
adverse effects that were systematically considered. The medi-
did not receive the complete prescribed 7 days of therapy with
an number of days (quartile range) in which trembling was ob-
the study medication. There were eight major deviations in
served was 0 (0–2) for albuterol and 0 (0) for placebo. The
the albuterol group and two in the placebo group. In the al-
median number of days (quartile range) in which the infant
buterol group, five infants stopped the study medication be-
vomited was 1.0 (0–3) for albuterol and 1.0 (0–2) for placebo.
Blinding was tested on study day 14 with a question re-
garding suspected treatment allocation. Both the primarycaregiver and the research nurse responded similarly: “don’tknow,” 60%; correct guess, 25%; and incorrect guess, 15%.
As in the single-dose studies,8,9 we found that extend-
ed-use oral albuterol was not effective. There were no practi-cal benefits in terms of the time to ROI or any of the specificsymptoms. This study was designed to mimic the customaryclinical use of albuterol in this patient group. Meaningful out-comes for parents and clinicians were chosen. The consistencyof the results in the primary and secondary outcomes empha-sizes the lack of difference between the course of illness in thetwo treatment groups.
These results add to the growing body of evidence that
suggests that bronchodilators, whether administered orally orby nebulization, are not effective therapies for infants withacute viral bronchiolitis. Two recent meta-analyses20,21 haveshown that the benefits of nebulized bronchodilators in thispatient group are limited. Some have suggested that bron-chodilators may be of benefit only in the very early stages ofillness, before epithelial necrosis predominates. Clinically itremains difficult to identify affected infants at this stage be-cause the first symptoms of bronchiolitis are common to manyviral infections. In otherwise healthy infants, we currently donot have the tools to predict, at the prodromal stage, whichinfants will develop bronchiolitis, nor the severity of illness.
Families seek healthcare when their affected infants developthe symptoms of respiratory distress; by then, the pathologicalpulmonary changes have already begun.
In this study, there was adequate power to show the in-
tended 2-day difference in the time to illness resolution, amagnitude of difference defined a priori as having clinical rel-evance. The time to ROI was measured by using the time ofstudy enrollment as a zero point. In fact, infants in both treat-ment groups were ill for a median of 4 days before ED presen-tation, as reported by parents. This prodromal duration isconsistent with the natural history of bronchiolitis with typi-cal onset of coryza, cough, and low-grade fever for several daysbefore more severe respiratory distress develops.
For practical purposes, adherence with allocated therapy
was measured by parental report. The numbers of singlemissed doses and late doses were similar between groups.
Overall reported adherence with the study medication washigh, regardless of allocation group. Interestingly, there weremore major protocol deviations in the albuterol group than inthe placebo group, and although five infants discontinued thetreatment, another three were actually started on off-study al-buterol by another physician. Of the five who discontinued
Survival curve results (see separate print-out of graph). A,
treatment, three did so because of negative side effects of al-
time to ROI; B,
time to normal feeding; C,
time to resolution of
time to resolution of coryza; E,
time to quiet breathing.P
values indicate significance by log-rank test.
Given budgetary constraints, a study nurse was available
to recruit infants between 8:00 AM and 12:00 AM and between6:00 PM and 11:00 PM during the study period. These times
ED were evaluated then. Nonetheless, 339 infants who were
were chosen because pilot data indicated that the majority of
potentially eligible were seen outside study hours. These in-
infants with mild-to-moderate bronchiolitis presenting to the
fants were similar in distribution by age and sex to infants in
Randomized, Double-Blind, Placebo-Controlled Trial of Oral Albuterolin Infants with Mild-to-Moderate Acute Viral Bronchiolitis
the study group and did not appear to be systematically differ-
Hanson IC, Shearer WT. Bronchiolitis. In: Oski FA, DeAngelis CD,
ent from infants evaluated during study hours. The 102 in-
Feigin RD, et al, editors. Principles and practice of pediatrics. 2nd ed.
fants who were excluded because of lack of informed consent
Philadelphia (PA): JB Lippincott Co; 1994. p. 1457-9.6.
Barone MA, editor. The Harriet Lane handbook. 14th ed. St Louis
were also similar to those in the study group. The majority of
the 42 ED physician refusals for patient participation were
Nahata MC, Schad PA. Pattern of drug usage in bronchiolitis. J Clin
linked to four physicians who thought that randomization and
blinded therapy were not appropriate in this patient group
Gadomski AM, Aref GH, El Din OB, El Sawy IH, Khallaf N, Black
based on their personal experiences. In accordance with the
RE. Oral versus nebulized albuterol in the management of bronchiolitis inEgypt. J Pediatr 1994;124:131-8.
guidelines provided by our ethics review board, physicians’
Gadomski AM, Lichenstein R, Horton L, King J, Keane V, Permutt T.
wishes for nonparticipation in the trial were respected.
Efficacy of albuterol in the management of bronchiolitis. Pediatrics
As the lack of effectiveness of bronchodilators has be-
come increasingly established,20–22 other therapies have been
Lowell DI, Lister G, Von Koss H, McCarthy P. Wheezing in infants:
considered in infants with bronchiolitis. Nebulized epineph-
the response to epinephrine. Pediatrics 1987;79:939-45.11.
Menon K, Sutcliffe T, Klassen TP. A randomized trial comparing the
rine has been used with some short-term benefits in acutely ill
efficacy of epinephrine with salbutamol in the treatment of acute bronchioli-
infants,11 but one recent trial found no benefits to extended
use in the hospitalized population with moderate to severe
Reijonen T, Korppi M, Pitkakangas S, Tenhola S, Remes K. The clinical
symptoms.22 No other effective therapies have been described
efficacy of nebulized epinephrine and albuterol in acute bronchiolitis. Arch
for the large population of previously well, first-time wheezing
Pediatr Adolesc Med 1995;149:686-92.13.
Klassen TP, Rowe PC, Sutcliffe T, Ropp LJ, McDowell IW, Li MM.
infants with mild-to-moderate viral bronchiolitis. Thus, in this
Randomized trial of salbutamol in acute bronchiolitis. J Pediatr 1991;118:
group of infants, supportive therapy remains most appropriate.
No significant group differences were found in the
Schweich PJ, Hurt TL, Walkley EI, Mullen N, Archibald LF. The
short-term symptom resolution in previously well infants with
use of nebulized albuterol in wheezing infants. Pediatr Emerg Care
mild-to-moderate bronchiolitis treated with oral albuterol
Klassen TP, Sutcliffe T, Watters LK, Wells GA, Allen UD, Li MM.
versus placebo. We do not recommend the use of oral albuterol
Dexamethasone in salbutamol-treated inpatients with acute bronchiolitis: a
randomized, controlled trial. J Pediatr 1997;130:191-6.16.
Cruz MN, Stewart G, Rosenberg N. Use of dexamethasone in the
We thank Michael B. Smith, Natalie Ross, and Patricia Vandercruys
outpatient management of acute laryngotracheitis.
for their valuable help with this trial.
Geelhoed GC, Turner J, Macdonald WBG. Efficacy of a small single
dose of oral dexamethasone for outpatient croup: a double-blind placebo-con-
Hanvey L, Avard D, Graham I, Underwood K, Campbell J, Kelly C.
Friedman LM, Furberg CD, De Mets DL. Fundamentals of clinical tri-
The health of Canada’s children. 2nd ed. Ontario, Ottawa (Canada): Canadi-
als. 2nd ed. Littleton (MA): PSG Publishing Co; 1985.
Landis JR, Koch GG. The measurement of observer agreement for cat-
Green M, Brayer AF, Schenkman KA, Wald ER. Duration of hospital-
egorical data. Biometrics 1977;33:159-74.
ization in previously well infants with respiratory syncytial virus infection. Pe-
Kellner JD, Gadomski AM, Ohlsson A, Wang EEL. Efficacy of bron-
chodilator therapy in bronchiolitis: a meta-analysis. Arch Pediatr Adolesc
Law BJ, De Carvalho V, The Pediatric Investigators Collaborative Net-
work on Infections in Canada. Respiratory syncytial virus infections in hospi-
Flores G, Horwitz RI. Efficacy of β2-agonists in bronchiolitis: a reap-
talized Canadian children: regional differences in patient populations and
praisal and meta-analysis. Pediatrics 1997;100:233-9.
management practices. Pediatr Infect Dis J 1993;12:659-63.
Patel H, Platt RW, Pekeles GS, Ducharme FM. A randomized, con-
Schweich P. Lower respiratory tract infections. In: Schwartz MW,
trolled trial of the effectiveness of nebulized therapy with epinephrine com-
Charney EB, Canny TA, et al, editors. Pediatric primary care. 2nd ed. Chica-
pared with albuterol and saline in infants hospitalized for acute viral
go (IL): Year Book Medical Publishers, Inc; 1990. p. 502-4.
bronchiolitis. J Pediatr 2002;141:818-24.
Murder and Restoration in 1950s Rural Minnesota: State v. Viola Gavle On Thursday morning, April 6, 1950, as Oscar Rasmussen sat with his friends Albert Knutson and Truman Gavle in a bar in Emmons, Minnesota – a tiny town a few miles south of Albert Lea, near the Iowa border – he wasn’t planning to die in the next hour. Instead, he and his two friends had a few beers and then decided
Newsletter 43 Gerald Mitchell We are very pleased to report that Gerald has made a remarkable recovery. He is at home, driving again and continuing to regain his strength. We send Ana and Gerald best wishes and hope to see him back on the croquet courts next season where he has been sorely missed. Results 2005 Season This has been another very successful season for the club, culminat