new england

The
journal of medicine
established in 1812 May 17, 2018 vol. 378 no. 20

Inhaled Combined Budesonide–Formoterol as Needed

in Mild Asthma
Paul M. O’Byrne, M.B., J. Mark FitzGerald, M.D., Eric D. Bateman, M.D., Peter J. Barnes, M.D., Nanshan Zhong, Ph.D.,
Christina Keen, M.D., Carin Jorup, M.D., Rosa Lamarca, Ph.D., Stefan Ivanov, M.D., Ph.D., and Helen K. Reddel, M.B., B.S., Ph.D.

a bs t r ac t

BACKGROUND
In patients with mild asthma, as-needed use of an inhaled glucocorticoid plus a fast- From the Firestone Institute for Respira-
acting β2-agonist may be an alternative to conventional treatment strategies. tory Health, St. Joseph’s Healthcare and
Department of Medicine, Michael G. De-
METHODS Groote School of Medicine, McMaster
University, Hamilton, ON (P.M.O.), and
We conducted a 52-week, double-blind trial involving patients 12 years of age or older the Institute for Heart and Lung Health,
with mild asthma. Patients were randomly assigned to one of three regimens: twice- University of British Columbia, Vancouver
daily placebo plus terbutaline (0.5 mg) used as needed (terbutaline group), twice-daily (J.M.F.) — both in Canada; the Division
of Pulmonology, Department of Medi-
placebo plus budesonide–formoterol (200 µg of budesonide and 6 µg of formoterol) cine, University of Cape Town, Cape Town,
used as needed (budesonide–formoterol group), or twice-daily budesonide (200 µg) plus South Africa (E.D.B.); Airway Disease Sec-
terbutaline used as needed (budesonide maintenance group). The primary objective tion, National Heart and Lung Institute,
Imperial College, London (P.J.B.); State
was to investigate the superiority of as-needed budesonide–formoterol to as-needed Key Laboratory of Respiratory Diseases,
terbutaline with regard to electronically recorded weeks with well-controlled asthma. First Affiliated Hospital, Guangzhou Med-
ical University, Guangzhou, China (N.Z.);
RESULTS AstraZeneca Research and Development,
A total of 3849 patients underwent randomization, and 3836 (1277 in the Gothenburg, Sweden (C.K., C.J., S.I.);
terbutaline group, 1277 in the budesonide–formoterol group, and 1282 in the AstraZeneca Research and Development,
Barcelona (R.L.); and Woolcock Institute
budesonide maintenance group) were included in the full analysis and safety data of Medical Research, University of Syd-
sets. With respect to the mean percentage of weeks with well-controlled asthma ney, Sydney (H.K.R.). Address reprint re-
per patient, budesonide–formoterol was superior to terbutaline (34.4% vs. 31.1% quests to Dr. O’Byrne at the Firestone In-
stitute for Respiratory Health, St. Joseph’s
of weeks; odds ratio, 1.14; 95% confidence interval [CI], 1.00 to 1.30; P = 0.046) Healthcare and Department of Medicine,
but inferior to budesonide maintenance therapy (34.4% and 44.4%, respectively; McMaster University, Rm. 2E1, 1280 Main
odds ratio, 0.64; 95% CI, 0.57 to 0.73). The annual rate of severe exacerbations was St. West, Hamilton, ON L8S 4K1, Canada,
or at [email protected].
0.20 with terbutaline, 0.07 with budesonide–formoterol, and 0.09 with budesonide
maintenance therapy; the rate ratio was 0.36 (95% CI, 0.27 to 0.49) for budesonide– N Engl J Med 2018;378:1865-76.
DOI: 10.1056/NEJMoa1715274
formoterol versus terbutaline and 0.83 (95% CI, 0.59 to 1.16) for budesonide– Copyright © 2018 Massachusetts Medical Society.
formoterol versus budesonide maintenance therapy. The rate of adherence in the
budesonide maintenance group was 78.9%. The median metered daily dose of inhaled
glucocorticoid in the budesonide–formoterol group (57 µg) was 17% of the dose in
the budesonide maintenance group (340 µg).
CONCLUSIONS
In patients with mild asthma, as-needed budesonide–formoterol provided superior
asthma-symptom control to as-needed terbutaline, assessed according to electronically
recorded weeks with well-controlled asthma, but was inferior to budesonide maintenance
therapy. Exacerbation rates with the two budesonide-containing regimens were similar
and were lower than the rate with terbutaline. Budesonide–formoterol used as needed
resulted in substantially lower glucocorticoid exposure than budesonide maintenance
therapy. (Funded by AstraZeneca; SYGMA 1 ClinicalTrials.gov number, NCT02149199.)
n engl j med 378;20 nejm.org May 17, 2018 1865
The New England Journal of Medicine
Downloaded from nejm.org on June 11, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

M
ild asthma, which can be well Symbicort Turbuhaler, AstraZeneca) used as need-
controlled either with reliever medica- ed, as compared with terbutaline (0.5 mg; terbu-
tion (short-acting β2-agonists [SABAs]) taline Turbuhaler, AstraZeneca) used as needed
used alone as needed or with low-dose inhaled and with twice-daily budesonide (200 µg; Pul-
glucocorticoid or leukotriene-receptor antagonist micort Turbuhaler, AstraZeneca) plus terbutaline
A Quick Take
is available at used as maintenance controller medication,1 oc- (0.5 mg) used as needed (Fig. 1). The trial sites
NEJM.org curs in approximately 50 to 75% of patients with are listed in the Supplementary Appendix, avail-
asthma.2 Symptoms may not necessarily be bur- able with the full text of this article at NEJM.org.
densome, but airway inflammation is usually The trial protocol, with the statistical analysis
present,3 and patients with mild asthma remain plan, is available at NEJM.org. The trial design
at risk for severe exacerbations (which account has been published previously.15
for 30 to 40% of asthma exacerbations leading
to emergency care2) and asthma-related death.2 Patients
Guidelines recommend that most adults and Patients, 12 years of age or older, who had re-
adolescents with asthma use regular daily low- ceived a clinical diagnosis of asthma (Global
dose inhaled glucocorticoids as maintenance Initiative for Asthma [GINA] 2012 criteria16) at
treatment to reduce airway inflammation, symp- least 6 months previously were eligible if they
toms, and the risk of exacerbations.1,4 However, had been assessed by the investigator as needing
in clinical practice, poor adherence to asthma GINA step 2 treatment16 for the 30 days before
medications, particularly inhaled glucocorticoids visit 2. Step 2 treatment is considered to be ap-
as maintenance therapy, is a major problem propriate in patients with asthma that is uncon-
across all severities of asthma,4-7 leading to under- trolled while the patient is taking inhaled short-
treatment of underlying inflammation and to an acting bronchodilators on an as-needed basis
increased risk of exacerbations.8-10 In parallel, pa- (subgroup 1 in our trial) or asthma that is well
tients rely on SABAs for symptom relief. However, controlled while the patient is taking mainte-
SABAs do not address the underlying inflamma- nance therapy with a low-dose inhaled glucocorti-
tory process or protect against exacerbations; coid or leukotriene-receptor antagonist plus short-
indeed, increased use of SABAs is associated acting bronchodilators used as needed (subgroup
with a higher exacerbation risk.11,12 2). Recruited patients were stratified according
One potential strategy to address these issues to pretrial treatment. Confirmation of the asthma
is the use of a combination of a fast-acting β2- diagnosis was required, either by a documented
agonist and an inhaled glucocorticoid taken only history of reversible airway obstruction or by
on an as-needed basis. This approach has proved means of a bronchodilator reversibility test con-
effective with beclomethasone and SABAs in ducted at visit 2 or 3 with an increase in the
patients with mild asthma13 and those with forced expiratory volume in 1 second (FEV1) of at
mild-to-moderate asthma.14 The objectives of the least 12% and 200 ml from the value obtained
Symbicort Given as Needed in Mild Asthma before bronchodilator use. Details of the inclu-
(SYGMA) 1 trial were to assess, among patients sion and exclusion criteria and stratification
with mild asthma, the long-term efficacy and technique are provided in the Supplementary
safety of budesonide–formoterol used as needed, Appendix.
measured according to electronically recorded The trial was performed in accordance with
weeks with well-controlled asthma and the rate the Declaration of Helsinki and Good Clinical
of severe exacerbations, as compared with terbu- Practice guidelines, and the protocol was approved
taline used as needed or budesonide maintenance by relevant authorities (Table S1 in the Supple-
therapy. mentary Appendix). All the patients provided
written informed consent (for patients younger
than 18 years of age, written informed consent
Me thods
was also obtained from a parent or guardian).
Trial Design
We conducted a double-blind, randomized, parallel- Trial Treatment
group, 52-week, phase 3 trial that evaluated the Before randomization, to confirm the appropri-
efficacy and safety of budesonide–formoterol ateness of GINA step 2 treatment,16 eligible pa-
(200 µg of budesonide and 6 µg of formoterol; tients entered a run-in period lasting 2 to 4 weeks

1866 n engl j med 378;20 nejm.org May 17, 2018

The New England Journal of Medicine

Downloaded from nejm.org on June 11, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 Budesonide–Formoterol as Needed in Mild Asthma

Randomization

Placebo twice per day+terbutaline as needed

Terbutaline
Placebo twice per day+budesonide–formoterol as needed
as needed
Budesonide twice per day+terbutaline as needed

Period Enrollment Run-in Treatment

Visit 1 2 3 4 5 6 7 8 Follow-
Trial entry Baseline up
Week −4 to −2 0 4 16 28 40 52 54

Electronic diary Electronic diary and inhaler monitor

and inhaler
monitor

Figure 1. Trial Design.

The terbutaline dose used during the run-in period (0.5 mg) corresponded to a delivered dose of 0.4 mg of terbutaline, delivered by a
Turbuhaler during the double-blind phase for blinding purposes. Pretrial asthma treatments were discontinued at visit 2. In order for
patients to be eligible to undergo randomization, morning and evening data must have been recorded for at least 8 days (any 8) of the
previous 10 days of the run-in period. The dose of budesonide–formoterol during the treatment period corresponded to a delivered
dose of 160 µg of budesonide and 4.5 µg of formoterol. An inhaler monitor recorded terbutaline use during the run-in period as well
as the use of each blinded trial inhaler. An electronic diary recorded the morning and evening peak expiratory flow, asthma symptoms,
and nighttime awakenings due to asthma and prompted the use of the blinded maintenance inhaler. Follow-up was conducted by means
of a telephone call.

during which they received only terbutaline on an of additional inhaled glucocorticoids was re-
as-needed basis (Fig. 1). To progress to random- corded.
ization (visit 3), patients must have used terbuta- Use of all trial medications or placebo during
line on an as-needed basis on at least 3 days the double-blind period and of terbutaline dur-
during the last week of the run-in period but ing the run-in period was recorded electroni-
could not have used six or more inhalations of cally with the use of an inhaler monitor (Turbu-
terbutaline per day for 2 or more days of 14 days haler usage monitor, Adherium).17 An electronic
in the run-in period (or for ≥3 days of 15 to 21 diary was used to record the morning and eve-
days or for ≥4 days of ≥22 days in the run-in ning peak expiratory flow, asthma symptoms,
period). Patients were also required to use the and nighttime awakenings due to asthma, and
trial-medication inhaler device and the electronic prompted use of the blinded maintenance inhaler.
diary correctly.
Patients were randomly assigned to one of End Points and Assessments
three regimens: twice-daily placebo plus terbuta- The primary objective was to show that
line (0.5 mg, used on an as-needed basis; terbu- budesonide–formoterol used as needed was su-
taline group); twice-daily placebo plus budesonide– perior to terbutaline used as needed in terms of
formoterol (200 µg of budesonide and 6 µg asthma symptom control, measured according
of formoterol, used on an as-needed basis; to the electronically recorded weeks with well-
budesonide–formoterol group); or twice-daily controlled asthma (see the Supplementary Appen-
budesonide (200 µg) plus terbutaline (0.5 mg, dix). This measurement was based on as-needed
used on an as-needed basis; budesonide mainte- use (according to the inhaler-monitor data),
nance group). During the trial, patients who had electronic-diary data for asthma symptom scores
asthma exacerbations or long-term poor asthma (scores were assessed on a 4-point scale ranging
control were permitted to receive additional treat- from 0 to 3, with higher values indicating more
ment with open-label budesonide at a dose of severe asthma symptoms), nighttime awakenings,
200 µg twice daily for 2 to 4 weeks or longer, and morning peak expiratory flow, and data
at the investigator’s discretion. The prescription from an electronic case-report form for the ad-

n engl j med 378;20 nejm.org May 17, 2018 1867

The New England Journal of Medicine
Downloaded from nejm.org on June 11, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

ditional use of inhaled or systemic glucocorti- sion. Writing and editing assistance, including
coids. A week could not be classified with well- preparation of a draft manuscript under the di-
controlled asthma unless the electronic diary rection and guidance of the authors, the incor-
was completed for at least 5 days, but a week poration of author feedback, and manuscript
could be classified with asthma being not well submission, was provided by inScience Commu-
controlled with as little as 1 day of data. nications, Springer Healthcare (funded by the
Secondary objectives included showing the sponsor), and by the sponsor.
noninferiority of budesonide–formoterol used as
needed to budesonide maintenance therapy with Statistical Analysis
regard to electronically recorded weeks with The sample size was estimated at 3750 patients
well-controlled asthma and comparing the rates (see the Supplementary Appendix). We estimated
and time to the first severe exacerbation (de- that 625 patients per treatment group and per
fined as worsening asthma leading to the use of subgroup according to pretrial treatment would
systemic glucocorticoids for ≥3 days, inpatient provide the trial with at least 95% power to com-
hospitalization, or an emergency department pare budesonide–formoterol used as needed with
visit leading to the use of systemic glucocorti- terbutaline used as needed, assuming an odds
coids) and the rates and time to the first moder- ratio of 1.39 between twice-daily budesonide
ate-to-severe exacerbation (including worsen- plus as-needed terbutaline and terbutaline used
ing asthma requiring the addition of inhaled as needed with regard to the electronically re-
budesonide at a dose of 200 µg twice daily to corded weeks with well-controlled asthma and
avoid progression to a severe exacerbation) in assuming that budesonide–formoterol used as
the budesonide–formoterol group versus the needed would have the same level of efficacy as
terbutaline group and versus the budesonide twice-daily budesonide. Testing was carried out
maintenance group. The descriptions of other at a two-sided alpha level of 0.05. In addition,
secondary efficacy end points, including Asthma the sample size allowed for 90% power to estab-
Control Questionnaire–5 (ACQ-5) scores, lung- lish noninferiority with regard to the electroni-
function variables, and quality of life (according cally recorded weeks of well-controlled asthma
to the Asthma Quality of Life Questionnaire with budesonide–formoterol used as needed
[AQLQ] score), have been published previously.15 as compared with twice-daily budesonide plus
The ACQ-5 consists of 5 questions about asthma as-needed terbutaline, with a prespecified non-
symptoms during the previous week, each scored inferiority limit of 0.8 (i.e., noninferiority was
on a range from 0 (no impairment) to 6 (maxi- concluded if the lower limit of the two-sided
mum impairment); the minimal clinically impor- 95% confidence interval of the odds ratio for
tant difference is 0.5 units. The AQLQ contains budesonide–formoterol used as needed, as com-
32 questions about asthma-related symptoms pared with twice daily budesonide plus terbuta-
and limitations during the preceding 2 weeks. line, was ≥0.8).
Each item is scored on a scale of 1 (severely im- The primary variable, electronically recorded
paired) to 7 (no impairment); the minimal clini- weeks with well-controlled asthma, was analyzed
cally important difference is 0.5 units. Safety by a repeated measures logistic-regression model
was evaluated according to the type, incidence, with treatment, pretrial treatment, and geograph-
and severity of adverse events and by monitoring ic region as fixed effects, and with trial week as
of vital signs. a categorical time variable. The model used an
exchangeable correlation structure. Odds ratios
Trial Oversight averaged over the 52-week period and their cor-
Trial data were collected by the clinical investi- responding 95% confidence intervals were de-
gators and were analyzed by employees of the rived from the model. The primary treatment
sponsor, AstraZeneca. The first and third au- comparison was budesonide–formoterol used as
thors vouch for the accuracy and completeness needed versus terbutaline used as needed (supe-
of the data and analyses and for the fidelity of riority test; the primary objective), and the sec-
the trial to the protocol. All the authors helped ondary comparison was budesonide–formoterol
draft each stage of the manuscript and read and used as needed versus budesonide maintenance
approved the final version at the time of submis- therapy (noninferiority test; the secondary ob-

1868 n engl j med 378;20 nejm.org May 17, 2018

The New England Journal of Medicine

Downloaded from nejm.org on June 11, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 Budesonide–Formoterol as Needed in Mild Asthma

jective). A hierarchical testing procedure was [CI], 1.00 to 1.30; P = 0.046). Thus, the odds of
performed, testing first the comparison of having a week with well-controlled asthma dur-
budesonide–formoterol used as needed versus ing the 52-week trial period were 14% higher in
terbutaline used as needed and then moving to the budesonide–formoterol group than in the
test budesonide–formoterol used as needed ver- terbutaline group.
sus twice-daily budesonide plus as-needed terbu-
taline if the result of the preceding test was Secondary Efficacy Outcomes
significant. Details of the analyses of the pri- Electronically Recorded Weeks with Well-Controlled
mary outcome, secondary outcomes, and superi- Asthma
ority and noninferiority testing are provided in Budesonide–formoterol used as needed was in-
the Supplementary Appendix. There was no ad- ferior to budesonide maintenance therapy with
justment for multiplicity testing of secondary regard to the percentage of electronically record-
variables. ed weeks with well-controlled asthma per pa-
tient (34.4% vs. 44.4%; odds ratio, 0.64; 95% CI,
0.57 to 0.73). The treatment effect was similar in
R e sult s
subgroup 1 and subgroup 2 (Fig. S2 in the Sup-
Patients plementary Appendix). Time-course results for
The trial was conducted from July 2014 through the electronically recorded weeks with well-
August 2017. Of the 5721 patients who were controlled asthma overall are shown in Figure 2,
enrolled, 3849 underwent randomization: 1280 and the individual components are shown in
patients were assigned to the terbutaline group, Figure S3 and Table S3 in the Supplementary
1279 to the budesonide–formoterol group, and Appendix. A prespecified analysis of the elec-
1290 to the budesonide maintenance group (Fig. tronically recorded weeks with well-controlled
S1 in the Supplementary Appendix). Overall, asthma, with removal of the “as-needed” com-
3836 patients had data that could be evaluated ponent, showed a decreased difference in the
for the full analysis and safety data sets, and treatment effect of budesonide maintenance ther-
3363 patients (87.4%) completed the trial. apy versus budesonide–formoterol used as need-
The demographic and clinical characteristics ed, from 36% to 22% (Table S4 in the Supple-
of the patients at baseline are shown in Table 1, mentary Appendix). Post hoc analysis of a
and in Table S2 in the Supplementary Appendix. modified end point of the electronically recorded
At trial entry, participants had uncontrolled weeks with well-controlled asthma, in which the
asthma symptoms (mean ACQ-5 score, 1.54) and first two inhalations used as needed per day
a mean bronchodilator reversibility of 15.4%. were not counted (i.e., were included as if they
Airflow limitation was mild (mean baseline FEV1 had been taken as maintenance doses), showed
before bronchodilator use, 84% of the predicted no difference between the budesonide–formoterol
value). In the year preceding enrollment, 19.7% group and the budesonide maintenance group
of the patients had had a severe exacerbation. (Table S5 in the Supplementary Appendix).
The treatment groups were well balanced, with
no clinically relevant differences in the baseline Exacerbations and Asthma-Related Discontinuations
characteristics. The subgroups according to pre- Budesonide–formoterol used as needed resulted
trial treatment had similar characteristics at in a 64% lower rate of severe exacerbations than
baseline, except that patients in subgroup 2 had terbutaline used as needed (annualized exacer-
slightly higher lung function than those in sub- bation rate, 0.07 vs. 0.20; rate ratio, 0.36; 95%
group 1. CI, 0.27 to 0.49) (Table 2, and Fig. S4 in the
Supplementary Appendix). The rates of severe
Primary Efficacy Outcome exacerbations in the budesonide–formoterol group
Budesonide–formoterol used as needed was su- and the budesonide maintenance group did not
perior to terbutaline used as needed with regard differ significantly (annualized exacerbation rate,
to the primary outcome of the mean percentage 0.07 and 0.09, respectively; rate ratio, 0.83; 95%
of electronically recorded weeks with well-con- CI, 0.59 to 1.16). Budesonide–formoterol used
trolled asthma per patient (34.4% vs. 31.1% of as needed also resulted in a 60% lower rate of
weeks; odds ratio, 1.14; 95% confidence interval moderate-to-severe exacerbations than terbuta-

n engl j med 378;20 nejm.org May 17, 2018 1869

The New England Journal of Medicine
Downloaded from nejm.org on June 11, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 1870
Table 1. Demographic and Clinical Characteristics of the Patients at Baseline, According to Treatment Group.*

Terbutaline Budesonide–Formoterol Budesonide Maintenance

as Needed as Needed Therapy Total
Characteristic (N = 1277) (N = 1277) (N = 1282) (N = 3836)

Age — yr 40.0±16.3 39.8±16.9 39.0±16.7 39.6±16.6

Female sex — no. (%) 771 (60.4) 777 (60.8) 797 (62.2) 2345 (61.1)
Downloaded from nejm.org on June 11, 2018. For personal use only. No other uses without permission.

Time since asthma diagnosis — yr

Median 6.3 6.5 6.3 6.4
Range 0.5–62.4 0.4–65.7 0.5–57.1 0.4–65.7
ACQ-5 score†
Copyright © 2018 Massachusetts Medical Society. All rights reserved.

Mean score
At trial entry 1.52±0.96 1.57±0.97 1.53±0.97 1.54±0.97
At baseline 1.54±0.95 1.61±0.97 1.55±0.96 1.57±0.96

The
Score ≥1.5 — no./total no. (%)

n e w e ng l a n d j o u r na l
The New England Journal of Medicine

At trial entry‡ 549/1160 (47.3) 601/1174 (51.2) 568/1177 (48.3) 1718/3511 (48.9)
n engl j med 378;20

At baseline 602/1256 (47.9) 649/1257 (51.6) 596/1257 (47.4) 1847/3770 (49.0)

AQLQ score§ 5.25±0.99 5.20±1.01 5.27±1.01 5.24±1.00
FEV1 — % of predicted value
Before bronchodilator use 84.13±14.08 84.18±14.24 84.23±13.91 84.18±14.07
After bronchodilator use 95.27±13.53 95.86±14.02 95.67±13.43 95.60±13.66
Peak expiratory flow ≥80% of the predicted value every morning 362/1276 (28.4) 340/1277 (26.6) 376/1282 (29.3) 1078/3835 (28.1)
nejm.org

— no./total no. (%)¶

Bronchodilator reversibility — % 14.4±11.5 14.9±11.3 14.6±11.6 14.6±11.5
Asthma control according to pretrial treatment — no. (%)∥
May 17, 2018

Uncontrolled with short-acting bronchodilator alone 565 (44.2) 565 (44.2) 576 (44.9) 1706 (44.5)

of
m e dic i n e
Controlled with inhaled glucocorticoid or leukotriene-receptor 712 (55.8) 712 (55.8) 706 (55.1) 2130 (55.5)
antagonist
Severe exacerbation in previous 12 mo — no. (%) 256 (20.0) 257 (20.1) 241 (18.8) 754 (19.7)

* Plus–minus values are means ±SD. There were no significant between-group differences in the demographic or clinical characteristics at baseline. Baseline was defined as the
assessment at visit 3 (i.e., the point at which randomization took place). FEV1 denotes forced expiratory volume in 1 second.
† The Asthma Control Questionnaire (ACQ-5) consists of five questions about asthma symptoms during the previous week, each of which is scored on a range from 0 (no impairment) to
6 (maximum impairment); the minimal clinically important difference is 0.5 units. Trial entry was defined as the assessment at the visit before the run-in period (i.e., visit 1 or 2). Data
at trial entry were missing for 117 patients in the terbutaline group, for 103 in the budesonide–formoterol group, and for 105 in the budesonide maintenance group; and data at baseline
were missing for 21, 20, and 25 patients, respectively.
‡ These calculations for data at trial entry were performed post hoc.
§ The standardized version of the Asthma Quality of Life Questionnaire (AQLQ) contains 32 questions about asthma-related symptoms and limitations during the preceding 2 weeks.
Each item is scored on a scale of 1 (severely impaired) to 7 (no impairment at all); the minimal clinically important difference is 0.5 units.
¶ Peak expiratory flow at this level was defined as a morning peak expiratory flow of at least 80% of the predicted value on every day of the previous 10 days in the run-in period.
∥ Control of asthma by the pretrial treatment was assessed by the physician.
 Budesonide–Formoterol as Needed in Mild Asthma

line used as needed (0.14 vs. 0.36), but the rate

in the budesonide–formoterol group did not dif- 100
Budesonide maintenance
90
fer significantly from that in the budesonide

Well-Controlled Asthma (%)

Budesonide–formoterol as needed
80 Terbutaline as needed

Patients with Week of

maintenance group (rate ratio, 0.95; 95% CI,
70
0.74 to 1.21) (Table 2, and Fig. S4 in the Supple- 60
mentary Appendix). 50
Budesonide–formoterol used as needed pro- 40
longed the time to the first severe exacerbation, 30
as compared with terbutaline used as needed 20
(hazard ratio, 0.44; 95% CI, 0.33 to 0.58). The 10
0
results in the budesonide–formoterol group did 0 4 8 12 16 20 24 28 32 36 40 44 48 52
not differ significantly from those in the Trial Treatment Week
budesonide maintenance group (hazard ratio,
0.90; 95% CI, 0.65 to 1.24) (Fig. 3). More patients Figure 2. Overall Weeks of Well-Controlled Asthma, According to Data
in the terbutaline group had asthma-related dis- in the Electronic Diary.
continuations than did those in the budesonide–
formoterol group or the budesonide maintenance
group (1.6% vs. 0.3% and 0.5%, respectively). 17% of that in the budesonide maintenance
The hazard ratio for the risk of asthma-related group (metered dose, 57 µg and 340 µg, respec-
discontinuation in the trial was 0.18 (95% CI, tively) (Table S6 in the Supplementary Appen-
0.06 to 0.52) in the budesonide–formoterol group dix). The total number of days with systemic
versus the terbutaline group and 0.66 (95% CI, glucocorticoid treatment for asthma was 465 days
0.19 to 2.35) in the budesonide–formoterol group in the budesonide–formoterol group, 500 days
versus the budesonide maintenance group (Fig. in the budesonide maintenance group, and 1237
S5 in the Supplementary Appendix). days in the terbutaline group.

Adherence and Glucocorticoid Dose Asthma-Control Questionnaire and Lung Function

Adherence to the twice-daily, blinded mainte- There were differences in the change from base-
nance regimen did not differ significantly across line in the ACQ-5 score in favor of the budeso-
the trial groups: the mean (±SD) percentage of nide–formoterol group versus the terbutaline
doses taken was 79.0±23.3% in the terbutaline group (mean difference, −0.15; 95% CI, −0.20 to
group, 79.1±23.0% in the budesonide–formoterol −0.11) and in favor of the budesonide mainte-
group, and 78.9±22.4% in the budesonide main- nance group versus the budesonide–formoterol
tenance group. Similar rates of adherence were group (mean difference, 0.15; 95% CI, 0.10 to
seen with the electronic diary. 0.20) (Table S7 in the Supplementary Appendix).
Additional inhaled or systemic glucocorticoids Similarly, there were differences between the
for asthma were prescribed in fewer patients re- budesonide–formoterol group and the other two
ceiving budesonide–formoterol as needed (12.8%) groups with regard to the average change from
than in those receiving terbutaline as needed baseline in the FEV1 before bronchodilator use
(27.0%) or budesonide maintenance therapy (mean change from baseline, 65.0 ml [95% CI,
(14.6%). The time to the use of additional gluco- 47.6 to 82.4] in the budesonide–formoterol group
corticoids for asthma was shorter in the terbuta- vs. 11.2 ml [95% CI, −6.4 to 28.9] in the terbu-
line group than in the budesonide–formoterol taline group and 119.3 ml [95% CI, 101.9 to
group (hazard ratio in the terbutaline group, 136.7] in the budesonide maintenance group)
0.41; 95% CI, 0.34 to 0.50); the time did not dif- (Table S8 in the Supplementary Appendix).
fer significantly between the budesonide main-
tenance group and the budesonide–formoterol Adverse Events
group (hazard ratio in the budesonide mainte- Adverse events were more frequent in the terbu-
nance group, 0.87; 95% CI, 0.70 to 1.07) (Fig. S6 taline group (in 545 of 1277 patients [42.7%])
in the Supplementary Appendix). than in the budesonide–formoterol group (485
The median daily dose of inhaled glucocorti- of 1277 [38.0%]) or the budesonide maintenance
coid in the budesonide–formoterol group was group (512 of 1282 [39.9%]) (Table S9 in the

n engl j med 378;20 nejm.org May 17, 2018 1871

The New England Journal of Medicine
Downloaded from nejm.org on June 11, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

Table 2. Summary of Asthma Exacerbations, According to Treatment Group.

Terbutaline Budesonide–Formoterol Budesonide Maintenance

as Needed as Needed Therapy
Variable (N = 1277) (N = 1277) (N = 1282)
All severe exacerbations
Patients with ≥1 exacerbation — no. (%) 152 (11.9) 71 (5.6) 78 (6.1)
Total no. of exacerbations 188 77 89
Annualized exacerbation rate 0.20 0.07 0.09
Comparison between as-needed budesonide–
formoterol and other regimen
Rate ratio 0.36 — 0.83
95% CI 0.27–0.49 — 0.59–1.16
P value <0.001 — 0.28
Severe exacerbation leading to hospitalization
Patients with ≥1 exacerbation — no. (%) 15 (1.2) 6 (0.5) 8 (0.6)
Total no. of exacerbations 21 6 8
Severe exacerbation leading to emergency department
visit and systemic glucocorticoid use
Patients with ≥1 exacerbation — no. (%) 29 (2.3) 7 (0.5) 10 (0.8)
Total no. of exacerbations 29 8 10
Severe exacerbation leading to systemic glucocorticoid
use for ≥3 days
Patients with ≥1 exacerbation — no. (%) 141 (11.0) 70 (5.5) 74 (5.8)
Total no. of exacerbations 173 76 84
All moderate or severe exacerbations
Patients with ≥1 exacerbation — no. (%) 274 (21.5) 131 (10.3) 143 (11.2)
Total no. of exacerbations 372 164 170
Annualized exacerbation rate 0.36 0.14 0.15
Comparison between as-needed budesonide–
formoterol and other regimen
Rate ratio 0.40 — 0.95
95% CI 0.32–0.49 — 0.74–1.21
P value <0.001 — 0.66

Supplementary Appendix). There were no nota- group (upper gastrointestinal hemorrhage and
ble differences in the adverse-event profile be- brain neoplasm, in 1 patient each) (Table S10 in
tween treatments, except that more adverse events the Supplementary Appendix).
led to discontinuation in the terbutaline group
(37 patients [2.9%]) than in the budesonide– Other Secondary End Points
formoterol group (10 patients [0.8%]) or the The results for the other secondary end points,
budesonide maintenance group (15 patients including peak expiratory flow values, symptom
[1.2%]). The number of patients with at least one and control scores, nighttime awakenings due
severe exacerbation leading to hospitalization to asthma, and medication use, are reported in
was greater in the terbutaline group (15 patients Tables S11 through S19 and Figures S7 and S8
[1.2%]) than in the budesonide–formoterol group in the Supplementary Appendix. The numbers of
(6 patients [0.5%]) or the budesonide mainte- patients with high use (>8 and >12 inhalations
nance group (8 patients [0.6%]) (Table 2). There in 1 day) of as-needed medication are reported
were two deaths in the budesonide maintenance in Table S20 in the Supplementary Appendix.

1872 n engl j med 378;20 nejm.org May 17, 2018

The New England Journal of Medicine

Downloaded from nejm.org on June 11, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 Budesonide–Formoterol as Needed in Mild Asthma

Terbutaline as needed Budesonide–formoterol as needed Budesonide maintenance

(N=1277) (N=1277) (N=1282)

A Severe Exacerbation
1.00 Budesonide–formoterol vs. terbutaline, P<0.001
0.25

Probability of Having a Severe Exacerbation

Budesonide–formoterol vs. budesonide, P=0.52

0.20
0.75
0.15

0.10
0.50

0.05

0.25 0.00
0 4 8 12 16 20 24 28 32 36 40 44 48 52

0.00
0 4 8 12 16 20 24 28 32 36 40 44 48 52
Weeks
No. at Risk
Terbutaline as needed 1277 1237 1190 1153 1131 1102 1084 1067 1038 1024 1017 987 977 731
Budesonide–formoterol as needed 1277 1258 1235 1218 1207 1179 1172 1159 1138 1127 1119 1097 1086 822
Budesonide maintenance 1282 1264 1238 1226 1201 1172 1159 1150 1136 1123 1110 1088 1076 811

B Moderate or Severe Exacerbation

0.25 Budesonide–formoterol vs. terbutaline, P<0.001

1.00 Budesonide–formoterol vs. budesonide, P=0.44

0.20
Probability of Having a Moderate

0.75 0.15
or Severe Exacerbation

0.10

0.50
0.05

0.00
0.25 0 4 8 12 16 20 24 28 32 36 40 44 48 52

0.00
0 4 8 12 16 20 24 28 32 36 40 44 48 52
Weeks
No. at Risk
Terbutaline as needed 1277 1210 1143 1098 1069 1031 1010 990 955 934 923 888 877 660
Budesonide–formoterol as needed 1277 1252 1227 1204 1184 1142 1130 1116 1089 1078 1067 1040 1028 778
Budesonide maintenance 1282 1257 1224 1206 1175 1143 1125 1111 1089 1074 1057 1031 1017 763

Figure 3. Time to First Exacerbation.

P values were not controlled for multiple comparisons. Insets show the same data on an enlarged y axis.

Discussion budesonide–formoterol used as needed was su-

perior to terbutaline used as needed for both
This trial showed that budesonide–formoterol symptom control, measured according to the
used as needed was a more effective treatment percentage of electronically recorded weeks with
than a SABA alone in patients with mild asthma; well-controlled asthma per patient, and the pre-

n engl j med 378;20 nejm.org May 17, 2018 1873

The New England Journal of Medicine
Downloaded from nejm.org on June 11, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

vention of moderate-to-severe and severe exacer- tory reliever approach that leverages patients’
bations. Although budesonide–formoterol used inherent relief-seeking behavior to also deliver
as needed was equally effective as budesonide inhaled glucocorticoids as soon as symptoms ap-
maintenance therapy in preventing moderate- pear, which provides a window of opportunity18-20
to-severe exacerbations, budesonide–formoterol that reduces the likelihood of progression to an
used as needed was inferior to budesonide main- exacerbation. Previous trials involving patients
tenance therapy in achieving electronically re- with moderate-to-severe asthma using mainte-
corded weeks with well-controlled asthma but nance and reliever therapy,21-27 involving patients
exposed the patients to less than one fifth of the with mild asthma using separate regimens,13 and
amount of inhaled glucocorticoid. involving patients with moderate asthma using
In interpretation of the comparisons of budeso- combination14 as-needed inhaled glucocorticoid
nide–formoterol used as needed with budesonide plus a SABA have also shown the advantages of
maintenance therapy, an important consideration this approach in reducing exacerbations and
is the extent to which the primary end point of maintaining symptom control at a lower total
the percentage of electronically recorded weeks dose of glucocorticoid.21-28
with well-controlled asthma per patient was The results of this trial also suggest that the
driven by the as-needed medication component. as-needed use of budesonide–formoterol in mild
Conventionally, symptoms and reliever use are asthma could address patients’ concerns about
both included in guideline-assessed symptom the risks of treatment, another issue that causes
control1 because, independent of symptoms, a overreliance on SABAs and poor adherence to
higher use of SABAs is associated with an in- maintenance treatment with an inhaled gluco-
creased exacerbation risk, which indicates a corticoid.11 Patients are often more concerned
greater need for preventive therapy. When the about adverse effects of inhaled glucocorticoids,7,29
reliever is a combined inhaled glucocorticoid even when low inhaled doses are used, than
plus β2-agonist, the amount used also represents their health care providers, and conversely they
the amount of preventive therapy that has been are less concerned about their level of symptom
delivered. Prespecified removal of the “as-needed” control.18,30 Since budesonide–formoterol used as
component from the definition of electronically needed was as effective as budesonide mainte-
recorded weeks with well-controlled asthma im- nance therapy in reducing exacerbation risk,
proved the treatment effect of budesonide–for- without the need for regular, twice-daily treat-
moterol used as needed versus both terbutaline ment, and resulted in only 17% of the inhaled
used as needed and budesonide maintenance glucocorticoid load, it would probably be accept-
therapy; however, the results still favored budeso- able to patients who have this concern and fits
nide maintenance therapy. with patients’ behavior.
In addition to day-to-day symptom control, The strengths of this trial include the 1-year
overall asthma control also includes the minimi- duration; the electronic monitoring of medica-
zation of the risk of adverse outcomes, including tion use, symptoms, and lung function; and the
exacerbations and adverse effects of medica- freedom to add open-label inhaled glucocorticoid
tions.1 The exacerbation rates in the terbutaline to avoid imbalance of withdrawals. The trial was
group in this trial showed that patients with designed to satisfy regulatory requirements for
mild asthma were at risk for exacerbations. The efficacy studies, and the high observed rate of ad-
facts that severe exacerbations and even death herence, approaching 80% with twice-daily re-
occur in patients with mild asthma,2 who repre- minders, means that budesonide maintenance
sent approximately 50 to 75% of patients with therapy was being evaluated under appropriate
asthma,2 and that 19.7% of the patients who conditions. Whether the results will be more favor-
underwent randomization in our trial reported able with budesonide–formoterol used as needed in
having had a severe exacerbation in the previous real-world populations in which adherence rates are
year, provide clinical relevance to the substantial considerably lower31 is currently being explored in
reduction in exacerbations achieved with budeso- ongoing studies (Australian New Zealand Clinical
nide–formoterol used as needed as compared Trials Registry numbers, ACTRN12615000999538
with terbutaline used as needed. We think that and ACTRN12616000377437).32,33
this finding is explained by the antiinflamma- One feature of this trial is the derivation of

1874 n engl j med 378;20 nejm.org May 17, 2018

The New England Journal of Medicine

Downloaded from nejm.org on June 11, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 Budesonide–Formoterol as Needed in Mild Asthma

the weeks with well-controlled asthma from the budesonide–formoterol used as needed. This
twice-daily electronically recorded diary, reliever relationship has been explored in the SYGMA 2
use, and peak expiratory flow; this approach trial (the results of which are reported in this
avoided retrospective data entry by patients and issue of the Journal37), which used a more prag-
may have resulted in a higher rate of reporting matic design to compare budesonide–formoterol
of symptoms, awakenings, and reliever use than used as needed with budesonide maintenance
has occurred in earlier studies in which patients therapy.
used paper-based diaries,34-36 thereby reducing In conclusion, this trial showed that budeso-
the overall percentage of electronically recorded nide–formoterol used as needed was superior to
weeks with well-controlled asthma. The double- the SABA terbutaline used as needed both for
blind, double-dummy design, although essential asthma symptom control and for reducing the
for showing the efficacy of a new regimen, risk of asthma exacerbations among patients
meant that patients who had been randomly as- with physician-assessed mild asthma. Further-
signed to the budesonide–formoterol group still more, budesonide–formoterol used as needed
had to use a twice-daily (placebo) inhaler, which was inferior to budesonide maintenance therapy
would not apply in clinical practice. These fac- with regard to electronically recorded weeks with
tors, together with the high rate of adherence well-controlled asthma but was similar to budeso-
to the maintenance regimen, may explain why nide maintenance therapy in reducing the risk of
budesonide–formoterol used as needed was infe- asthma exacerbations, at a substantially lower
rior to twice-daily budesonide maintenance ther- total glucocorticoid load and without the need
apy with regard to the electronically recorded for adherence to a twice-daily maintenance-
weeks with well-controlled asthma. Neverthe- therapy schedule.
less, the findings indicate that, in patients with
mild asthma who were able to maintain high Supported by AstraZeneca.
Disclosure forms provided by the authors are available with
adherence to twice-daily medication, regular low- the full text of this article at NEJM.org.
dose inhaled glucocorticoid remained more effec- We thank the health care providers, research staff, patients,
tive in achieving daily asthma control and equally and caregivers who participated in this trial; and Vicky Hinstridge,
David Candlish, Matt Weitz, and Amy Evans of inScience Com-
effective with respect to severe exacerbations, munications, Springer Healthcare, for medical writing assis-
albeit with greater glucocorticoid exposure, than tance with an earlier version of the manuscript.

References
1. 2018 GINA report, global Initiative concordance: implications for asthma ized controlled trial. JAMA 2012;308:987-
for Asthma (GINA). Global strategy for treatment. Chest 2006;130:Suppl:65S-72S. 97.
asthma management and prevention (http:// 8. Bårnes CB, Ulrik CS. Asthma and ad- 14. Papi A, Canonica GW, Maestrelli P, et
ginasthma.org/2018-gina-report-global herence to inhaled corticosteroids: cur- al. Rescue use of beclomethasone and albu-
-strategy-for-asthma-management-and rent status and future perspectives. Respir terol in a single inhaler for mild asthma.
-prevention/). Care 2015;60:455-68. N Engl J Med 2007;356:2040-52.
2. Dusser D, Montani D, Chanez P, et al. 9. Suissa S, Ernst P, Kezouh A. Regular 15. O’Byrne PM, FitzGerald JM, Zhong N,
Mild asthma: an expert review on epide- use of inhaled corticosteroids and the long et al. The SYGMA programme of phase 3
miology, clinical characteristics and treat- term prevention of hospitalisation for trials to evaluate the efficacy and safety of
ment recommendations. Allergy 2007;62: asthma. Thorax 2002;57:880-4. budesonide/formoterol given ‘as needed’
591-604. 10. Wu AC, Butler MG, Li L, et al. Primary in mild asthma: study protocols for two
3. Vignola AM, Chanez P, Campbell AM, adherence to controller medications for randomised controlled trials. Trials 2017;
et al. Airway inflammation in mild inter- asthma is poor. Ann Am Thorac Soc 2015; 18:12.
mittent and in persistent asthma. Am J 12:161-6. 16. Global Initiative for Asthma (GINA).
Respir Crit Care Med 1998;157:403-9. 11. O’Byrne PM, Jenkins C, Bateman ED. Global strategy for asthma management
4. Shahidi N, FitzGerald JM. Current rec- The paradoxes of asthma management: and prevention, 2012 (http://ginasthma
ommendations for the treatment of mild time for a new approach? Eur Respir J .org/).
asthma. J Asthma Allergy 2010;3:169- 2017;50:1701103. 17. Pilcher J, Shirtcliffe P, Patel M, et al.
76. 12. Stanford RH, Shah MB, D’Souza AO, Three-month validation of a turbuhaler
5. Jónasson G, Carlsen KH, Mowinckel P. Dhamane AD, Schatz M. Short-acting electronic monitoring device: implications
Asthma drug adherence in a long term clin- β-agonist use and its ability to predict for asthma clinical trial use. BMJ Open
ical trial. Arch Dis Child 2000;83:330-3. future asthma-related outcomes. Ann Respir Res 2015;2(1):e000097.
6. Jónasson G, Carlsen KH, Sødal A, Allergy Asthma Immunol 2012;109:403-7. 18. Partridge MR, van der Molen T, Myrs-
Jonasson C, Mowinckel P. Patient com- 13. Calhoun WJ, Ameredes BT, King TS, eth SE, Busse WW. Attitudes and actions
pliance in a clinical trial with inhaled et al. Comparison of physician-, biomarker-, of asthma patients on regular mainte-
budesonide in children with mild asthma. and symptom-based strategies for adjust- nance therapy: the INSPIRE study. BMC
Eur Respir J 1999;14:150-4. ment of inhaled corticosteroid therapy in Pulm Med 2006;6:13.
7. Horne R. Compliance, adherence, and adults with asthma: the BASALT random- 19. Bousquet J, Boulet L-P, Peters MJ, et al.

n engl j med 378;20 nejm.org May 17, 2018 1875

The New England Journal of Medicine
Downloaded from nejm.org on June 11, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.
 Budesonide–Formoterol as Needed in Mild Asthma

Budesonide/formoterol for maintenance et al. Efficacy and safety of budesonide/ Bryan S. The burden of nonadherence
and relief in uncontrolled asthma vs. formoterol single inhaler therapy versus a among adults with asthma: a role for
high-dose salmeterol/fluticasone. Respir higher dose of budesonide in moderate to shared decision-making. Allergy 2017;72:
Med 2007;101:2437-46. severe asthma. Curr Med Res Opin 2004; 705-12.
20. Buhl R, Kuna P, Peters MJ, et al. The 20:1403-18. 32. Beasley R, Pavord I, Papi A, et al. De-
effect of budesonide/formoterol mainte- 26. Patel M, Pilcher J, Pritchard A, et al. scription of a randomised controlled trial
nance and reliever therapy on the risk of Efficacy and safety of maintenance and re- of inhaled corticosteroid/fast-onset LABA
severe asthma exacerbations following liever combination budesonide-formoterol reliever therapy in mild asthma. Eur
episodes of high reliever use: an explor- inhaler in patients with asthma at risk of Respir J 2016;47:981-4.
atory analysis of two randomised, con- severe exacerbations: a randomised con- 33. Fingleton J, Hardy J, Baggott C, et
trolled studies with comparisons to stan- trolled trial. Lancet Respir Med 2013;1: al. Description of the protocol for the
dard therapy. Respir Res 2012;13:59. 32-42. PRACTICAL study: a randomised con-
21. Demoly P, Louis R, Søes-Petersen U, 27. Papi A, Corradi M, Pigeon-Francisco C, trolled trial of the efficacy and safety of
et al. Budesonide/formoterol maintenance et al. Beclometasone-formoterol as main- ICS/LABA reliever therapy in asthma. BMJ
and reliever therapy versus conventional tenance and reliever treatment in patients Open Respir Res 2017;4(1):e000217.
best practice. Respir Med 2009;103:1623- with asthma: a double-blind, randomised 34. Ferrari R, Russell AS. Effect of a
32. controlled trial. Lancet Respir Med 2013; symptom diary on symptom frequency
22. Kuna P, Peters MJ, Manjra AI, et al. 1:23-31. and intensity in healthy subjects. J Rheu-
Effect of budesonide/formoterol mainte- 28. Rabe KF, Pizzichini E, Ställberg B, et al. matol 2010;37:2387-9.
nance and reliever therapy on asthma ex- Budesonide/formoterol in a single inhaler 35. Hufford MR, Stone AA, Shiffman S,
acerbations. Int J Clin Pract 2007;61:725- for maintenance and relief in mild-to- Schwartz JE, Broderick JE. Paper vs. elec-
36. moderate asthma: a randomized, double- tronic diaries: compliance and subject
23. O’Byrne PM, Bisgaard H, Godard PP, blind trial. Chest 2006;129:246-56. evaluations. Applied Clinical Trials. Au-
et al. Budesonide/formoterol combination 29. Boulet LP. Perception of the role and gust 2002:38-43.
therapy as both maintenance and reliever potential side effects of inhaled cortico- 36. Bateman ED, Boushey HA, Bousquet J,
medication in asthma. Am J Respir Crit steroids among asthmatic patients. Chest et al. Can guideline-defined asthma con-
Care Med 2005;171:129-36. 1998;113:587-92. trol be achieved? The Gaining Optimal
24. Rabe KF, Atienza T, Magyar P, Larsson 30. Price D, Fletcher M, van der Molen T. Asthma ControL study. Am J Respir Crit
P, Jorup C, Lalloo UG. Effect of budesonide Asthma control and management in Care Med 2004;170:836-44.
in combination with formoterol for reliev- 8,000 European patients: the REcognise 37. Bateman ED, Reddel HK, O’Byrne PM,
er therapy in asthma exacerbations: a ran- Asthma and LInk to Symptoms and Expe- et al. As-needed budesonide–formoterol
domised controlled, double-blind study. rience (REALISE) survey. NPJ Prim Care versus maintenance budesonide in mild
Lancet 2006;368:744-53. Respir Med 2014;24:14009. asthma. N Engl J Med 2018;378:1877-87.
25. Scicchitano R, Aalbers R, Ukena D, 31. Pollard S, Bansback N, FitzGerald JM, Copyright © 2018 Massachusetts Medical Society.

ARTICLE METRICS NOW AVAILABLE

Visit the article page at NEJM.org and click on Metrics to view comprehensive and
cumulative article metrics compiled from multiple sources, including Altmetrics.
www.nejm.org/about-nejm/article-metrics.

1876 n engl j med 378;20 nejm.org May 17, 2018

The New England Journal of Medicine

Downloaded from nejm.org on June 11, 2018. For personal use only. No other uses without permission.
Copyright © 2018 Massachusetts Medical Society. All rights reserved.