CHEST Original Research

COPD

Safety and Efficacy of Combined

Long-Acting ␤-Agonists and Inhaled
Corticosteroids vs Long-Acting
␤-Agonists Monotherapy for Stable COPD
A Systematic Review
Gustavo J. Rodrigo, MD; José A. Castro-Rodriguez, MD, PhD;
and Vicente Plaza, MD

Background: Current guidelines recommend the use of inhaled corticosteroids (ICSs) added to
long-acting ␤2-agonists (LABAs) for treatment of symptomatic patients with severe and very
severe COPD. However, the evidence has been inconclusive. The aim of this review was to assess
the safety and efficacy of LABAs/ICSs compared with LABA monotherapy for patients with
moderate-to-very severe COPD.
Methods: Systematic searches were conducted on MEDLINE, EMBASE, the Cochrane Con-
trolled Trials Register, and the trial registers of manufacturers, without language restriction.
Primary outcomes were COPD exacerbations and mortality. Secondary outcomes included lung
function, health-related quality of life, and adverse effects.
Results: Eighteen randomized controlled trials (12,446 participants) were selected. Therapy with
LABAs/ICSs did not decrease the number of severe exacerbations (relative risk [RR], 0.91; 95% CI,
0.82 to 1.01; I2 ⴝ 1%), or all-cause mortality (RR, 0.90; 95% CI, 0.76 to 1.06; I2 ⴝ 0%), respiratory
mortality (RR, 0.80; 95% CI, 0.61 to 1.05; I2 ⴝ 0%), and cardiovascular mortality (RR, 1.22; 95% CI,
0.88 to 1.71; I2 ⴝ 0%). To the contrary, the number of moderate exacerbations (RR, 0.84; 95% CI,
0.74 to 0.96; I2 ⴝ 50%) and the St. George respiratory questionnaire total score (weighted mean
difference, ⴚ1.88; 95% CI, ⴚ2.44 to ⴚ1.33; I2 ⴝ 29%) were significantly reduced with LABA/ICS
therapy. Although therapy with LABAs/ICSs increases FEV1 significantly (0.06 and 0.04 L, respec-
tively), they were associated with an increased risk of pneumonia (RR, 1.63; 95% CI, 1.35 to 1.98;
I2 ⴝ 20%).
Conclusions: Compared with LABA monotherapy, the magnitude of the benefits of LABA/ICS
therapy did not reach that of the criteria for predefined clinically important effects and were
associated with serious adverse effects. (CHEST 2009; 136:1029 –1038)
Abbreviations: ICS ⫽ inhaled corticosteroid; LABA ⫽ long-acting ␤2-agonist; MI ⫽ myocardial infarction; NNTB ⫽
number needed to treat for benefit; NNTH ⫽ number needed to treat for harm; RR ⫽ relative risk; SABA ⫽ short-acting
␤2-agonist; SGRQ ⫽ St. George respiratory questionnaire; WMD ⫽ weighted mean difference

C isOPDcharacterized
is a preventable and treatable disease that
by airflow limitation that is not
fully reversible.1,2 The main therapeutic goals are to
prevent and control symptoms, reduce the frequency
and severity of exacerbations, and improve health
status and exercise tolerance. Current guidelines1,2
recommend a stepwise increase in treatment, de-
pending on the severity of the disease. Short-acting
inhaled bronchodilators (short-acting ␤2-agonists
[SABAs] and anticholinergic agents) are recom-
mended for the relief of symptoms on an as-needed
basis, whereas long-acting inhaled bronchodilators
(long-acting ␤2-agonists [LABAs] or tiotropium) in
a regularly scheduled regimen are recommended
as first-line therapy in symptomatic patients with
moderate-to-very severe COPD.
Evidence3 shows that LABA monotherapy is asso-
ciated with significant improvements regarding
COPD exacerbations, pulmonary function, quality
of life, and use of rescue medication, with a low

www.chestjournal.org CHEST / 136 / 4 / OCTOBER, 2009 1029

Downloaded from chestjournal.chestpubs.org at Downey Community Hospital on July 14, 2010

© 2009 American College of Chest Physicians
incidence of adverse effects. However, in view of the
multicomponent nature of COPD, the use of inhaled
corticosteroids (ICSs), particularly in combination
with a LABA has obtained widespread acceptance
among clinicians. In fact, the addition of ICSs in
patients with severe or very severe disease (stages III
or IV) with repeated exacerbations is recommended
for decreasing exacerbation rates, and improving
lung function and health status.1,2 Nevertheless, the
evidence of the superiority of combination therapy
(LABAs/ICSs) over LABA monotherapy has been
inconclusive. While a 2007 review4 failed to demon-
strate the superiority of combination therapy over
LABA monotherapy in reducing COPD exacerba-
tions, others have reported5–7 some benefits of ther-
apy with LABAs/ICSs in terms of COPD exacerba-
tions, pulmonary function, and quality of life. Even
so, these conclusions might be questionable because
they are based on a reduced number of selected
studies and outcomes. Consequently, we performed
a systematic review to assess the safety and efficacy
of the use of LABAs/ICSs in COPD patients com-
pared with LABA monotherapy. The following two
specific questions were identified: (1) what are the
risks of adding an ICS to a LABA compared with
LABAs monotherapy? and (2) does therapy with
LABAs/ICSs provide significant clinical benefits
compared with LABA monotherapy?
Materials and Methods
Search Strategy and Selection Criteria
We identified studies from MEDLINE, EMBASE (January
1980 to May 2009), and the Cochrane Controlled Trials Register
(second quarter of 2009) databases by using the following MeSH,
full text, and keywords terms: (long-acting ␤2 adrenoceptor
agonist OR salmeterol OR formoterol OR inhaled corticoste-
roids OR fluticasone OR budesonide OR beclomethasone)
AND (COPD OR chronic bronchitis OR emphysema). Also,
we performed a search of relevant files from AstraZeneca
Manuscript received April 2, 2009; revision accepted July 2, 2009.
Affiliations: From the Departamento de Emergencia (Dr. Ro-
drigo), Hospital Central de las Fuerzas Armadas, Montevideo,
Uruguay; the School of Medicine (Dr. Castro-Rodriguez), Pon-
tificia Universidad Católica de Chile, Santiago, Chile; and Servei
de Pneumologia (Dr. Plaza), Hospital de la Santa Creu i Sant
Pau, Universitat Autonoma de Barcelona, Barcelona, Spain.
Funding/Support: The funding for this study came from salary
support for Dr. Rodrigo. No sponsorship from institutions or
pharmaceutical industry was provided to conduct this study.
Correspondence to: Gustavo J. Rodrigo, MD, Departamento
de Emergencia, Hospital Central de las Fuerzas Armadas, Av 8
de Octubre 3020, Montevideo 11600, Uruguay; e-mail: gurodrig@
adinet.com.uy
© 2009 American College of Chest Physicians. Reproduction
of this article is prohibited without written permission from the
American College of Chest Physicians (www.chestjournal.org/site/
misc/reprints.xhtml).
DOI: 10.1378/chest.09-0821
1030
Downloaded from chestjournal.chestpubs.org at Downey Community Hospital on July 14, 2010
© 2009 American College of Chest Physicians
(www.astrazenecaclinicaltrials.com) and GlaxoSmithKline (www.
gsk-clinicalstudyregister.com) databases. Trials published solely
in abstract form were excluded because the methods and results
could not be fully analyzed.
The specific inclusion criteria were as follows: (1) stable adult
patients aged ⬎ 40 years with COPD satisfying the diagnostic
criteria of the American Thoracic Society/European Respiratory
Society1 or the Global Initiative for Chronic Obstructive Lung
Disease2; (2) therapy with inhaled LABAs plus ICSs (delivered
via metered-dose inhaler or dry powder inhaler) as the interven-
tion arm compared with therapy with a LABA; (3) study dura-
tions of ⬎ 1 month; (4) randomized controlled trials (parallel-
group design) without language restriction; and (5) primary
outcomes “severe COPD exacerbation” (requiring hospitalization
or withdrawals) and “moderate COPD exacerbations” (requiring
systemic corticosteroids or antibiotic use), all-cause mortality
(deaths for any cause), respiratory deaths (deaths due to a
respiratory event such as COPD exacerbation or pneumonia),
and cardiovascular mortality (including sudden death) during the
treatment period. Secondary outcome measures were as follows:
mean change in FEV1 (pre-bronchodilator therapy and post-
bronchodilator therapy); mean change from baseline in the St.
George respiratory questionnaire (SGRQ) total score8; end-of-
treatment dyspnea score; withdrawals from the study during the
treatment period (overall, due to adverse effects, and due to lack
of efficacy); and adverse effects (pneumonia, oropharyngeal
candidiasis, viral respiratory infections, and myocardial infarc-
tions [MIs]).
Data Abstraction and Validity Assessment
Titles, abstracts, and citations were independently analyzed by
all reviewers. From full text, they independently assessed studies
for inclusion based on the criteria for population, intervention,
study design, and outcomes. Three reviewers (G.J.R., J.C.R., and
V.P.) were independently involved in all stages of study selection,
data extraction, and quality assessment. Any disagreement was
resolved by consensus. In case of multiple published or unpub-
lished reports for a particular study, data from the most recent
version were extracted.
Statistical Analysis
Binary outcomes were pooled by using common relative risks
(RRs) and 95% CIs. If pooled effect estimates were significantly
different between groups, we calculated the number needed to
treat for benefit (NNTB) or the number needed to treat for harm
(NNTH). For continuous outcomes, the standardized mean
difference or weighted mean difference (WMD) and 95% CIs
were calculated. Heterogeneity was further measured by using
the I2 test.9 With low heterogeneity (I2 ⬍ 40%), data were
combined by mean of a fixed-effects model10; otherwise, a
random-effects model was used. Publication bias of primary
outcomes was evaluated by means of the visual inspection of
funnel plots.11 A predefined sensitivity analysis of the primary
outcome of severe COPD exacerbations was conducted to ex-
plore the influence of the following factors: concealment alloca-
tion12 (adequate vs unclear); trial duration (long-term [ⱖ 52
weeks] vs short-term [⬍ 52 weeks]); reversibility to SABA (poorly
reversible patients or FEV1 ⬍ 15% from baseline vs reversible
patients or FEV1 ⱖ 15% from baseline); choice of LABAs
(salmeterol vs formoterol); and the use of ICSs before the
patients were enrolled (ⱖ 50% of patients vs ⬍ 50% of patients).
Subgroups were compared by using the interaction test.13 A
p value of ⬍ 0.05 using a two-tailed test was considered to
indicate significance. Metaanalyses were performed with using a
Original Research

Figure 1. Flowchart for identification of studies.
statistical software package (Review Manager, version 5.0.20; the
Nordic Cochrane Centre, the Cochrane Collaboration; Copen-
hagen, Denmark).
Results
Of 164 potential relevant citations, 18 randomized,
controlled trials14 –31 fulfilled the inclusion crit-
eria (Fig 1). Five trials were unpublished.27–31 Data
analysis was restricted to the LABAs/ICSs and
LABAs arms of those trials (12,446 patients) [Table
1]. Five studies used formoterol/budesonide combi-
nation therapy,17,20,26,27 and 13 studies used salmet-
erol/fluticasone combination therapy.14 –16,18,19,21–31
Seven studies14 –16,21,23,24,31 evaluated therapy with
inhaled fluticasone in combination with salmeterol
at a dosage of 500 ␮g twice daily, seven stud-
ies18,19,22,25,28 –30 assessed therapy with fluticasone at
a dosage of 250 ␮g twice daily, and four stud-
ies17,20,26,27 evaluated therapy with budesonide at a
dosage of 320 ␮g twice daily. All trials used single
inhalers containing both ICSs and LABAs to deliver the
combined therapy. Eleven trials15–21,23–26 reported that
a mean of 31% of patients (range, 0 to 55% of patients)
had received ICSs before they were enrolled in the
study. There were 11 long-term trials (ie, ⱖ 52
www.chestjournal.org
Downloaded from chestjournal.chestpubs.org at Downey Community Hospital on July 14, 2010
© 2009 American College of Chest Physicians
weeks)16 –18,20,21,23–25,27–29 and 7 short-term trials (ie,
⬍ 52 weeks).14,15,19,22,26,30,31 Studies enrolled mostly
stable patients with COPD who met the Global Initia-
tive for Chronic Obstructive Lung Disease criteria for
moderate-to-very severe COPD exacerbations.2 The
mean age of patients was 64 years (72% of patients
were male), with an average baseline FEV1 of 40% of
predicted normal values. Allocation concealment was
adequate in only 5 studies,16,20,21,23,24 and was unclear
in the remaining 13 studies.
Primary Outcomes
Compared with LABA monotherapy, combination
therapy with LABAs/ICSs did not significantly
decrease the risk of severe COPD exacerbations
(11.3% vs 12.5%, respectively) [Table 2]. The post
hoc subgroup analysis did not show significant dif-
ferences in COPD exacerbations regarding conceal-
ment allocation, trial duration, reversibility to SABA
use, LABA choice, and use of ICSs before the
patients were enrolled in the study (Table 3). To the
contrary, the use of LABAs/ICSs was associated with
a significantly reduced risk of moderate COPD
exacerbations when compared with LABAs alone
(17.5% vs 20.1%, respectively), with evidence of
statistical heterogeneity among trials. The NNTB
was 31 (95% CI, 20 to 93).
Patients receiving therapy with LABAs/ICSs were
not associated with a significant decrease of overall
mortality when compared with those receiving
LABAs alone (4.5% vs 5.5%, respectively) [Fig 2]. In
the same way, the metaanalysis did not show
significant differences between groups regarding
the risk of respiratory deaths (1.8% in the LABAs/
ICSs group vs 2.4% in the LABAs group) and
cardiovascular mortality (1.6% vs 1.4%, respectively),
without evidence of statistical heterogeneity (Fig 2).
On visual inspection of the funnel plots (Fig 3),
publication bias could be ruled out for all-cause,
respiratory, and cardiovascular mortality. To the
contrary, for severe COPD exacerbations, the plot
presented an asymmetrical shape with an absence of
small studies showing a group benefit for therapy
with LABAs.
Secondary Outcomes
Regarding pulmonary function, patients treated
with LABAs/ICSs showed significantly greater in-
creases in the mean change in FEV1 from baseline
(pre-bronchodilator therapy and post-bronchodilator
therapy) compared with patients treated with LABAs
alone (Table 4); however, both comparisons showed
statistical heterogeneity. Patients receiving therapy
with LABAs/ICSs showed a significantly greater
reduction in the SGRQ total score. Finally, at the
CHEST / 136 / 4 / OCTOBER, 2009 1031
 Table 1—Characteristics of Included Studies

Mean
Study Location/ Patients: No. Age, Mean Baseline Current Bronchodilator Intervention: Allocation
Studies Duration (% male) yr FEV1 Smoker Response, % Dose Concealment

Cazzola et al14 Single-center/12 wk S: 20 (90) 65 1.15 L 45 pack-yr ⬍ 12 S: 50 ␮g twice Unclear

MDI
SF: 40 (92) 64 1.16 L 44 pack-yr ⬍ 12 SF: 50/250 or
50/500 ␮g
twice MDI
Mahler et al15 69 centers/24 wk S: 160 (64) 64 40% predicted 46% 21 S: 50 ␮g twice Unclear
DPI
SF: 165 (62) 63 41% predicted 46% 21 SF: 50/500 ␮g
twice DPI
Calverley et al16 196 centers/52 wk S: 372 (70) 63 46% predicted 51% 3.7 S: 50 ␮g twice Adequate
DPI
SF: 358 (75) 63 49% predicted 52% 4.0 SF: 50/500 ␮g
twice DPI
Calverley et al17 109 centers/52 wk F: 255 (75) 63 36% predicted 36% 6 F: 9 ␮g twice Unclear
DPI
FB: 254 (78) 64 36% predicted 33% 6 FB: 9/320 ␮g
twice DPI
Dal Negro Single-center/52 wk S: 6 (100) 55–78 48% predicted 0% 3 S: 50 ␮g twice Unclear
et al18 DPI
SF: 5 (83) 53–77 50% predicted 0% 3.5 SF: 50/ 250 ␮g
twice DPI
Hanania et al19 75 centers/24 wk S: 177 (58) 64 42% predicted 51% 20 S: 50 ␮g twice Unclear
DPI
SF:178 (61) 63 41% predicted 43% 19 SF: 50/250 ␮g
twice DPI
Szafranski et al20 89 centers/52 wk F: 201 (76) 63 36% predicted 38% 6 F: 9 ␮g twice Adequate
DPI
FB:208 (76) 64 36% predicted 30% 6 FB: 9/320 ␮g
twice DPI
Wouters et al21 39 centers/52 wk S: 184 (75) 64 1.41 L 35% 4 S: 50 ␮g twice Adequate
DPI
SF: 189 (73) 63 1.41 L 39% 4 SF: 50/500 ␮g
twice DPI
O’Donnell 21 centers/8 wk S: 59 (75) 65 40% predicted 34% 18 S: 50 ␮g twice Unclear
et al22 DPI
SF: 62 (69) 63 42% predicted 42% 18 SF: 50/250 ␮g
twice DPI
Calverley et al23 444 centers/156 wk S: 1,521 (76) 65 44% predicted 43% 4 S: 50 ␮g twice Adequate
DPI
SF: 1,533 (75) 65 44% predicted 43% 4 SF: 50/500 ␮g
twice DPI
Kardos et al24 95 centers/ 44 wk S: 487 (78) 64 40% predicted 37% 7 S: 50 ␮g twice Adequate
DPI
SF: 507 (74) 64 40% predicted 37% 6 SF: 50/500 ␮g
twice DPI
Ferguson et al25 94 centers/52 wk S: 388 (52) 65 33% predicted 38% NA S: 50 ␮g twice Unclear
DPI
SF: 394 (58) 65 33% predicted 40% NA SF: 50/250 ␮g
DPI twice
Tashkin et al26 194 centers/26 wk F: 284 (66) 64 39% predicted 42% NA F: 9 ␮g twice Unclear
DPI
FB: 845 (69) 63 39% predicted 44% NA FB: 9/160 ␮g or
320/9 ␮g
twice MDI
D-5899C0 237 centers/52 wk F: 495 (65) 63 34% predicted 41% 16 F: 9 ␮g twice Unclear
000127 DPI
FB: 988 (63) 64 34% predicted 36% 16 FB: 9/320 ␮g or
160/9 MDI
twice
(Continued)

1032 Original Research

Downloaded from chestjournal.chestpubs.org at Downey Community Hospital on July 14, 2010

© 2009 American College of Chest Physicians
 Table 1—(Continued)

Mean
Study Location/ Patients: No. Age, Mean Baseline Current Bronchodilator Intervention: Allocation
Studies Duration (% male) yr FEV1 Smoker Response, % Dose Concealment

SCO4004128 31 centers/156 wk S: 94 (63) 66 ⬍ 70% predicted NA NA S: 50 ␮g twice Unclear

DPI
SF: 92 (63) 65 ⬍ 70% predicted NA NA SF: 50/250 ␮g
DPI twice
SCO10025029 98 centers/52 wk S: 403 (57) 65 ⬍ 50% predicted ⬎ 10 pack-yr NA S: 50 ␮g twice Unclear
DPI
SF: 394 (51) 65 ⬍ 50% predicted ⬎ 10 pack-yr NA SF: 50/250 ␮g
DPI twice
SCO10047030 135 centers/24 wk S: 532 (78) 64 1.68 L 44% NA S: 50 ␮g twice Unclear
DPI
SF: 518 (82) 64 1.65 L 42% NA SF: 50/250 ␮g
DPI twice
SCO10492531 11 centers/12 wk S: 38 (79) 64 ⬍ 50% predicted ⬎ 10 pack-yr NA S: 50 ␮g twice Unclear
DPI
SF: 39 (82) 64 ⬍ 50% predicted ⬎ 10 pack-yr NA SF: 50/500 ␮g
DPI twice
NA ⫽ not applicable; S ⫽ salmeterol; F ⫽ formoterol; SF ⫽ salmeterol/fluticasone; FB ⫽ formoterol/budesonide; MDI ⫽ metered-dose inhaler;
DPI ⫽ dry powder inhaler.

end of the protocol, patients treated with combina-
tion therapy had a significantly lower dyspnea score
compared with patients treated with LABAs alone,
although with evidence of statistical heterogeneity
among the trials (Table 4).
LABAs/ICSs therapy was associated with a signif-
icant decrease of overall withdrawals from the study
(Table 4). Furthermore, patients receiving therapy
with LABAs/ICSs were associated with a significantly
lower rate of withdrawals from the study due to lack
of efficacy, but not with significant withdrawals from
the study due to adverse effects (Table 4). However,
the use of LABAs/ICSs was also associated with
significantly increased rates of pneumonia (63%
increase in RR), viral respiratory infections (22%
increase in RR), and oropharyngeal candidiasis (59%
increase in RR) compared with the use of LABAs
alone. Finally, combined therapy did not show a
significant difference in the rate of MI, compared
with LABA monotherapy.
Discussion
In the present study, which is the largest system-
atic review designed to evaluate the safety and
efficacy of the regular use of LABAs/ICSs compared
with the use of LABAs alone in stable patients with
moderate-to-very severe COPD, we found that treat-
ment with LABAs/ICSs did not modify the risks of
overall mortality, respiratory deaths, and cardiovas-
cular mortality (primary outcomes) compared with

Table 2—Analysis of Primary Outcomes (LABAs/ICSs vs LABA)

LABAs ⫹ ICS,
No./Total No. LABAs, No./ Absolute Risk
Outcomes References (%) Total No. (%) Measure (95% CI) Reduction I2, %
COPD exacerbations 15–17,19–30 757/6,685 (11.3) 704/5,612 (12.5) RR ⫽ 0.91 (0.82–1.01) ⫺1.2 (⫺2.4 to 0.0) 1
(requiring
hospitalization or
withdrawal)
COPD exacerbations 15–26,29 794/4,532 (17.5) 1,015/5,058 (20.1) RR ⫽ 0.84 (0.74–0.96); ⫺2.5 (⫺1.0 to ⫺4.1) 50
requiring systemic p ⫽ 0.008/NNTB ⫽
corticosteroids 31 (20–93)
All-cause mortality 16,17,20,21, 240/5,292 (4.5) 261/4,721 (5.5) RR ⫽ 0.90 (0.76–1.06) ⫺1.0 (⫺1.8 to 0.0) 0
23–26,28–30
Respiratory deaths 16,17,20,21, 94/5,292 (1.8) 114/4,721 (2.4) RR ⫽ 0.80 (0.61–1.05) ⫺0.6 (⫺1.2 to 0.0) 0
23–26,28–30
Cardiovascular mortality 16,21,23–26, 72/5,856 (1.6) 63/5,299 (1.4) RR ⫽ 1.22 (0.88–1.71) 0.2 (⫺0.3 to 0.7) 0
28–30

www.chestjournal.org CHEST / 136 / 4 / OCTOBER, 2009 1033

Downloaded from chestjournal.chestpubs.org at Downey Community Hospital on July 14, 2010

© 2009 American College of Chest Physicians
 Table 3—Sensitivity Analysis. Comparisons Between RR in COPD Exacerbations Requiring Hospitalization or
Withdrawal) Stratified by Concealment Allocation (Adequate vs Unclear), Trial Duration (Long-Term > 52 Weeks
vs Short-Term < 52 Weeks), Reversibility to SABAs (Poorly Reversible Patients or FEV1 < 15% From Baseline vs
Reversible Patients or FEV1 > 15%), Baseline Severity (Moderate to Severe vs Severe to Very Severe), LABAs
choice (Salmeterol vs Formoterol), and Use of ICSs Before the Patients Were Enrolled (< 50% of Patients vs < 50%
of Patients)
Interactive Test17
Subgroup Comparisons RR (95% CI) 关I2兴 RR (95% CI) p Value

Adequate 16,20,21,23,24
vs unclear 0.93 (0.82–1.05) 关35%兴 vs 0.90 (0.77–1.05) 关0%兴 0.96 (0.79–1.18) 0.74
concealment14,15,17–19,22,25–31
Long-term16,17,20,21,23,25,27–29 vs 0.93 (0.84–1.03) 关1%兴 vs 0.80 (0.59–1.11) 关0%兴 0.86 (0.61–1.19) 0.37
short-term15,19,22,24,26,30
Poorly reversible16,17,20,21,23,24 vs 0.90 (0.80–1.01) 关0%兴 vs 0.94 (0.74–1.20) 关0%兴 1.04 (0.79–1.36) 0.75
reversible15,19,22,25–30
Salmeterol15,16,19,21–25,28–30 vs 0.94 (0.84–1.05) 关0%兴 vs 0.87 (0.73–1.04) 关41%兴 0.92 (0.75–1.14) 0.46
Formoterol17,20,26,27
Previous use of ICS (ⱖ 50%16, 24,26 vs 0.84 (0.65–1.14) 关28%兴 vs 0.92 (0.88–1.04), 关31%兴 1.06 (0.79–1.43) 0.65
⬍ 50% of patients15,17–21,23–25)

treatment with LABAs alone. On the contrary, the
analysis of secondary outcomes showed that therapy
with LABAs/ICSs significantly increased the risk of
pneumonia, oropharyngeal candidiasis, and viral re-
spiratory infections (question 1). However, it is
interesting to point out that these adverse effects
were not accompanied by a concomitant and propor-
tional increase in respiratory-related mortality or over-
all mortality. Concerning the benefits (question 2), we
found that therapy with LABAs/ICSs significantly
decreased the frequency of moderate COPD exac-
erbations independently of concealment allocation,
trial duration, reversibility to SABA therapy, LABA
choice, and previous use of ICSs.
In the same way, therapy with LABAs/ICSs was
associated with significant increases in the mean
change in pre-bronchodilator therapy and post-
bronchodilator therapy FEV1, the mean change in
SGRQ total score, and with a significant decrease in
the end-of-treatment dyspnea score compared with
treatment with LABAs alone. However, given that
the size of these benefits did not reach the suggested
clinically important minimal differences (FEV1, 0.10
to 0.14 L; SGRQ score, 4-unit decrement),32 the
relevance of these improvements seems uncertain.
In the same way, the 16% decrease in the rate of
moderate COPD exacerbations was smaller than the
suggested threshold value of 22% for clinical signif-
icance.32 Conversely, treatment with LABAs/ICSs
failed to significantly reduce the risk of severe
COPD exacerbations. These facts suggest a limited
extra efficacy when ICSs were added to LABAs for
COPD treatment.
The relative benefits of therapy with LABAs/ICSs
must be weighed against the risks. Thus, the most
concerning side effect was the increase in the risk of
pneumonia associated with the administration of
ICSs added to LABAs. The precise mechanism is
uncertain, but it could be related to the fact that
ICSs achieve locally high concentrations in the lung,
increasing the risk of pneumonia due to their immu-
nosuppressive effects.33 Thus, inhaled fluticasone at
dosages of 1,000 ␮g/d exerts effects on serum corti-
sol levels that are equivalent to 10 mg of prednisone,
a dose that may double the risk of pneumonia in
patients with arthritis.34 Our analysis showed an
increase in the risk of pneumonia with both moder-
ate dosages of fluticasone (500 ␮g/d; RR, 1.75; 95%
CI, 1.16 to 2.64; I2 ⫽ 30%) and high dosages of
fluticasone (1,000 ␮g/d; RR ⫽ 1.64; 95% CI, 1.32 to
2.06, I2 ⫽ 22%). To the contrary, therapy with
LABAs/ICSs was associated with significant de-
creases in overall withdrawals and withdrawals due
to lack of efficacy compared with LABA mono-
therapy. This finding could be associated with the
fact that therapy with LABAs/ICSs significantly re-
duces dyspnea with a greater clinical effectiveness
perception. Also, this fact could be associated with
better control of the disease.
When we compared the results of the present
metaanalysis with those of previous reviews (based
on a limited number of published trials), we found
some similarities and few differences. For example, a
Cochrane review5 of 10 studies, reported a signifi-
cant RR reduction of moderate COPD exacerbations
of 18% with the combined treatment with LABAs/
ICSs compared with treatment with LABAs alone.
Also, the combination therapy was more effective
than that with LABAs alone in improving quality
of life as measured by the SGRQ (⫺1.64 points),
and predose and postdose FEV1 (0.06 and 0.05 L,
respectively). While there was no significant dif-
ference in terms of overall mortality, pneumonia
occurred more commonly in patients receiving

1034 Original Research

Downloaded from chestjournal.chestpubs.org at Downey Community Hospital on July 14, 2010

© 2009 American College of Chest Physicians
 Figure 2. Pooled RR for overall mortality, respiratory deaths, and cardiovascular mortality (with 95%
CIs) of eligible studies comparing inhaled LABAs/ICSs with LABAs.

combined therapy (58% increase in RR). How-
ever, another metaanalysis4 that was limited to five
English-language trials failed to demonstrate the
superiority of combination treatment over LABA
monotherapy in reducing COPD exacerbations and
overall mortality.
More recently, Sobieraj et al7 on the basis of seven
studies reported that therapy with LABAs/ICSs de-
creased the risk of moderate COPD exacerbations
(relative reduction risk of 18%), decreased the
SGRQ total score (⫺1.98 points), and increased the
risk of pneumonia (32% increase in RR) compared
with therapy with LABAs alone. Additionally, com-
bined therapy also showed a reduced risk in overall
withdrawals (17% increase in RR), and without
difference in the rate of overall mortality between
groups. Another systematic review35 has found no
evidence that therapy with LABAs alone is more
effective than combined therapy with LABAs/ICSs.
In the same study, a subgroup analysis showed that,
when added to LABAs, ICSs significantly reduced
the number of exacerbations in patients with FEV1
ⱕ 40% predicted. However, these conclusions are
uncertain because the review was based on only

www.chestjournal.org CHEST / 136 / 4 / OCTOBER, 2009 1035

Downloaded from chestjournal.chestpubs.org at Downey Community Hospital on July 14, 2010

© 2009 American College of Chest Physicians
Figure 3. Funnel plots of LABAs/ICSs vs LABAs comparing COPD severe exacerbations (A), all-cause mortality (B), respiratory deaths (C),
and cardiovascular mortality (D).

seven studies15–17,19,20,23,24 (6,376 patients with
COPD). Finally, Singh et al36 showed a significantly
increased risk of pneumonia with combined therapy
compared with LABA monotherapy (32% increase in
RR) without a significantly increased risk of death.
Interestingly, the risk of pneumonia could be specif-
ically attributed to the use of ICSs because the risk
for pneumonia associated with ICS use was similar
when therapy with ICSs were compared with pla-
cebo or when ICSs were added to therapy with
LABAs and compared with LABA therapy alone. In
the same way, a recent trial37 comparing therapy
with inhaled tiotropium with therapy with salmet-
erol/fluticasone in patients with COPD showed an
increased hazard ratio for time to reported pneumo-
nia for combined therapy of ⬎ 2 years (94% increase
in RR).
This review was performed according to the meth-
odological criteria suggested for scientific reviews.38
The fact that there was low evidence of clinical and
statistical heterogeneity between studies increased
the confidence of our findings. Furthermore, there
was no evidence of publication bias in the majority of
primary outcomes. However, our metaanalysis had
several potential limitations that came from the
quality of the reported data. Thus, the trials did not
consistently use similar definitions of COPD exacer-
bations or pneumonia. In particular, we recognized
that the severity of exacerbations is a complex con-
cept constituted by several factors, and that this fact
could modify our results. Also, most of the studies
were not specifically designed to monitor outcomes
as all-cause, respiratory, or cardiovascular mortality.
Additionally, the risk of bias was unclear in 13 trials
in the analysis. Also, the fact that 80% of the
reviewed patients were men limits the applicability
of the results since COPD is suspected to affect men
and women equally.
This metaanalysis confirms and extends data from
previous reviews. The main results of our review are
as follows: LABAs/ICSs did not decrease the risk of
all-cause, respiratory, and cardiovascular mortality;
however, therapy with LABAs/ICSs increases the
risk of pneumonia, oropharyngeal candidiasis, and
viral respiratory infections. The use of LABAs/ICSs
was associated with a lower incidence of moderate
COPD exacerbations (but not of severe exacerba-
tions), increased pulmonary function, improved dys-

1036 Original Research

Downloaded from chestjournal.chestpubs.org at Downey Community Hospital on July 14, 2010

© 2009 American College of Chest Physicians
 Table 4 —Analysis of Secondary Outcomes (LABAs/ICSs vs LABAs)

LABAs ⫹ ICSs,
No./Total No. LABAs, No./ Absolute Risk
Outcomes References (%) Total No. (%) Measure (95% CI) 关p Value兴 Reduction I2, %
Mean change in 14–16,19,21–27, 5,613 5,082 WMD ⫽ 0.06 (0.04–0.07) 82
pre-bronchodilator 29–30 关0.0001兴
therapy FEV1, L
Mean change in 14–16,19,22, 3,455 2,501 WMD ⫽ 0.04 (0.02–0.05) 64
post-bronchodilator 23,26,27 关0.0001兴
therapy FEV1, L
Mean change in 16,17,20, 4,617 4,040 WMD ⫽ ⫺1.88 (⫺2.44 to ⫺1.33) 29
SGRQ 23–26,30 关0.0001兴
End-of-treatment 15,16,19,21,22, 3,216 2,643 SMD ⫽ ⫺0.20 (⫺0.25 to ⫺0.15) 82
dyspnea score 24–26,30 关0.0001兴
Total withdrawals 14–17,19–31 1,731/5,919 (29.2) 1,731/5,919 (29.2) RR ⫽ 0.87 (0.82–0.92) 关0.0001兴; ⫺1.6 (⫺0.2 to ⫺3.0) 0
NNTB ⫽ 25 (18–41)
Withdrawals due to 15,16, 19–21, 680/5,381 (12.6) 654/4,803 (13.6) RR ⫽ 0.93 (0.84–1.03) 关0.26兴 ⫺1.0 (⫺0.3 to 2.3) 0
adverse effects 23–26,28–30
Withdrawals due to 15,16,23,25, 63/3,362 (1.9) 117/3,376 (3.5) RR ⫽ 0.54 (0.40–0.72) 关0.0001兴; ⫺1.6 (⫺0.8 to ⫺2.4) 0
lack of efficacy 29,30 NNTB ⫽ 73 (56–120)
Pneumonia 15,16,19,21, 263/5,212 (5.0) 153/4,540 (3.4) RR ⫽ 1.63 (1.35–1.98) 关0.0001兴; 1.7 (0.9 to 2.5) 20
23–26,28–31 NNTH ⫽ 40 (26–72)
Oropharyngeal 16,19,21,22, 292/3,521 (8.4) 200/2,741 (7.2) RR ⫽ 1.59 (1.07–2.37) 关0.002兴; 1.2 (0.1 to 2.3) 65
candidiasis 24,26,27,29 NNTH ⫽ 22 (10–179)
Viral respiratory 15–17,19,22,23, 441/4,844 (9.1) 342/4,362 (7.8) RR ⫽ 1.22 (1.07–1.39) 关0.004兴; 1.3 (0.1 to 2.4) 0
infections 25,27,29,30 NNTH ⫽ 57 (33–179)
MI 16,23–25,28,29 34/3,278 (1.0) 33/3,265 (1.0) RR ⫽ 1.03 (0.64–1.64) 关0.91兴 0 (⫺0.5 to 0.5) 13
SMD ⫽ standardized mean difference.

pnea, and health-related quality-of-life total scores. Acknowledgments

However, the magnitude of these benefits did not
reach the recent predefined criteria to be clinical
important.32
Current guidelines1,2 recommend the use of
ICSs in combination with LABAs to reduce the
frequency of exacerbations in symptomatic pa-
tients with severe and very severe COPD. Further-
more, some authors39 have suggested that, in COPD
patients as in asthma patients, concomitant ICS
therapy is preferable over LABA monotherapy. Nev-
ertheless, this review suggests that combination
therapy with LABAs/ICSs presents a borderline
statistical and limited clinical significance compared
with LABA monotherapy. Moreover, combination
therapy offers no statistically significant additional
survival benefit and increased the risk of serious
adverse effects. Even so, this last issue requires
further prospective evaluation in large studies using
objective definitions of pneumonia. It is likely that
most patients with COPD with these levels of sever-
ity should be treated only with LABA monotherapy.
However, it is possible also that a future definition of
different COPD phenotypes will allow us to know
which patients can benefit from ICSs, and which
should only be treated with LABAs. Thus, patients
who benefit from combination therapy with LABAs/
ICSs could be those patients with steroid-responsive
eosinophilic bronchitis.
Author contributions: Dr. Rodrigo (1) has made substantial
contributions to conception and design, acquisition of data, and
analysis and interpretation of data; (2) has drafted the submitted
article and revised it critically for important intellectual content;
and (3) has provided final approval of the version of the article to
be published. Dr. Castro-Rodriguez (1) has made substantial
contributions to conception and design, and interpretation of
data; (2) has revised the article critically for important intellectual
content; and (3) has provided final approval of the version to be
published. Dr. Plaza (1) has made substantial contributions to
conception and design, and interpretation of data; (2) has revised
the article critically for important intellectual content; and (3) has
provided final approval of the version to be published.
Financial/nonfinancial disclosures: Dr. Rodrigo has partici-
pated as a lecturer and speaker in scientific meetings and courses
under the sponsorship of Boehringer Ingelheim, GlaxoSmithKline,
AstraZeneca, Dr. Esteve SA, and Merck Sharp and Dome. Dr.
Castro-Rodriguez has participated as a lecturer and speaker in
scientific meetings and courses under the sponsorship of Merck
Sharp and Dohme, GlaxoSmithKline, and Grunenthal; and as
member of the advisory board for GlaxoSmithKline. Dr. Plaza has
participated as a lecturer and speaker in scientific meetings and
courses under the sponsorship of AstraZeneca, GlaxoSmithKline,
Dr. Esteve SA, and Merck Sharp and Dohme.
References
1 Celli BR, MacNee W, ATS/ERS Task Force. Standards for the
diagnosis and treatment of patients with COPD: a summary of
the ATS/ERS position paper. Eur Respir J 2004; 23:932–946
2 Global Initiative for Chronic Obstructive Lung Disease. Global
strategy for the diagnosis, management and prevention of

www.chestjournal.org CHEST / 136 / 4 / OCTOBER, 2009 1037

Downloaded from chestjournal.chestpubs.org at Downey Community Hospital on July 14, 2010

© 2009 American College of Chest Physicians
 COPD. Available at: www.goldcopd.org. Accessed August 27,
2009
3 Rodrigo GJ, Nannini LJ, Rodriguez-Roisín R. Safety of
long-acting ␤-agonists in stable COPD: a systematic review.
Chest 2008; 133:1079 –1087
4 Wilt TJ, Niewoehner D, MacDonald R, et al. Management of
stable chronic obstructive pulmonary disease: a systematic
review for a clinical practice guideline. Ann Intern Med 2007;
147:639 – 653
5 Nannini LJ, Cates CJ, Lasserson TJ, et al. Combined corticoste-
roid and long-acting ␤-agonist in one inhaler versus long-acting
␤-agonists for chronic obstructive pulmonary disease. Cochrane
Database Syst Rev (database online). Issue 4, 2007
6 Hanania NA. The impact of inhaled corticosteroid and long-
acting ␤-agonist combination therapy on outcomes in COPD.
Pulm Pharmacol Ther 2008; 21:540 –550
7 Sobieraj DM, White M, Coleman CI. Benefits and risks of
adjunctive inhaled corticosteroids in chronic obstructive pulmo-
nary disease: a meta-analysis. Clin Ther 2008; 30:1416 –1425
8 Jones PW, Quirk FH, Baveystock CM. The St. George’s
respiratory questionnaire. Respir Med 1991; 85:25–31
9 Higgins JPT, Thompson SG, Deecks JJ, et al. DG. Measuring
inconsistency in meta-analyses. BMJ 2003; 327:557–560
10 Mantel N, Haenszel W. Statistical aspects of the analysis of
data from retrospective studies of disease. J Natl Cancer Inst
1959; 22:719 –748
11 Egger M, Davey Smith G, Schneider M, et al. Bias in
meta-analysis detected by a simple, graphical test. BMJ 1997;
315:629 – 634
12 Higgins JPT, Green S, eds. Cochrane handbook for system-
atic reviews of interventions version 5.0.1 [updated Septem-
ber 2008]. Available at: www.cochrane-handbook.org. Ac-
cessed August 27, 2009
13 Altman DG, Bland JM. Interaction revisited: the difference
between two estimates. BMJ 2003; 326:219
14 Cazzola M, Di Lorenzo G, Di Perna F, et al. Additive effects
of salmeterol and fluticasone or theophylline in COPD. Chest
2000; 118:1576 –1581
15 Mahler DA, Wire P, Horstman D, et al. Effectiveness of
fluticasone propionate and salmeterol combination delivered
via the diskus device in the treatment of chronic obstructive
pulmonary disease. Am J Respir Crit Care Med 2002; 166:
1084 –1091
16 Calverley P, Pauwels R, Vestbo J, et al. Combined salmeterol
and fluticasone in the treatment of chronic obstructive pul-
monary disease: a randomized controlled trial. Lancet 2003;
361:449 – 456
17 Calverley PM, Boonsawat W, Cseke Z, et al. Maintenance
therapy with budesonide and formoterol in chronic obstruc-
tive pulmonary disease. Eur Respir J 2003; 22:912–919
18 Dal Negro RW, Pomari C, Tognella S, et al. Salmeterol and
fluticasone 50 microg/250 microg bid in combination provides
a better long-term control than salmeterol 50 microg bid
alone and placebo in COPD patients already treated with
theophylline. Pulm Pharmacol Ther 2003; 16:241–246
19 Hanania NA, Darken P, Horstman D, et al. The efficacy and
safety of fluticasone propionate (250 ␮g)/salmeterol (50 ␮g)
combined in the discus inhaler for the treatment of COPD.
Chest 2003; 124:834 – 843
20 Szafranski A, Cukier A, Ramirez A, et al. Efficacy and safety
of budesonide/formoterol in the management of chronic
obstructive pulmonary disease. Eur Respir J 2003; 21:74 – 81
21 Wouters EFM, Postma DS, Fokkens B, et al. Withdrawal of
fluticasone propionate from combined salmeterol/fluticasone
1038
Downloaded from chestjournal.chestpubs.org at Downey Community Hospital on July 14, 2010
© 2009 American College of Chest Physicians
treatment in patients with COPD causes immediate and
sustained disease deterioration: a randomized controlled trial.
Thorax 2005; 60:480 – 487
22 O’Donnell DE, Sciurba F, Celli B, et al. Effect of fluticasone
propionate/salmeterol on lung hyperinflation and exercise
endurance in COPD. Chest 2006; 130:647– 656
23 Calverley PM, Anderson JA, Celli B, et al. Salmeterol and
fluticasone propionate and survival in chronic obstructive
pulmonary disease. N Engl J Med 2007; 356:775–789
24 Kardos P, Wencker M, Glaab T, et al. Impact of salmeterol/
fluticasone propionate versus salmeterol on exacerbations in
severe chronic obstructive pulmonary disease. Am J Respir
Crit Care Med 2007; 175:144 –149
25 Ferguson GT, Anzueto A, Fei R, et al. Effect of fluticasone
propionate/salmeterol (250/50 mg) or salmeterol (50 mg) on
COPD exacerbations. Respir Med 2008; 102:1099 –1108
26 Tashkin DP, Rennard SI, Martin P, et al. Efficacy and safety
of budesonide and formoterol in one pressurized metered-
dose inhaler in patients with moderate to very severe chronic
obstructive pulmonary disease: results of a 6-month random-
ized clinical trial. Drugs 2008; 68:1975–2000
27 AstraZeneca. Clinical trial register: D-5899C00001 trial. Avail-
able at: http://www.astrazenecaclinicaltrials.com/clinicaltrials/. Ac-
cessed August 27, 2009
28 GlaxoSmithKline. Clinical trial register: SCO40041 trial.
Available at: http://www.gsk-clinicalstudyregister.com/. Accessed
August 27, 2007
29 GlaxoSmithKline. Clinical trial register: SCO100250 trial. Avail-
able at: http://www.gsk-clinicalstudyregister.com/. Accessed
August 27, 2007
30 GlaxoSmithKline. Clinical trial register: SCO100470 trial. Avail-
able at: http://www.gsk-clinicalstudyregister.com/. Accessed
August 27, 2007
31 GlaxoSmithKline. Clinical trial register: SCO104925 trial.
Available at: http://www.gsk-clinicalstudyregister.com/. Ac-
cessed August 27, 2007
32 Cazzola M, MacNee W, Martinez FJ, et al. Outcomes for
COPD pharmacological trials: from lung function to biomar-
kers. Eur Respir J 2008; 31:416 – 468
33 Suissa S, McGhan R, Niewoehner D, et al. Inhaled cortico-
steroids in chronic obstructive pulmonary disease. Proc Am
Thorac Soc 2007; 4:535–542
34 Wolfe F, Caplan L, Michaud K. Treatment for rheumatoid
arthritis and the risk of hospitalization for pneumonia: asso-
ciations with prednisone, disease modifying antirheumatic
drugs, and anti-tumor necrosis factor therapy. Arthritis
Rheum 2006; 54:628 – 634
35 Puhan MA, Bachmann LM, Kleijnen J, et al. Inhaled drugs
to reduce exacerbations in patients with chronic obstruc-
tive pulmonary disease: a network meta-analysis. BMC
Med 2009; 7:2
36 Singh S, Amin AV, Loke YK. Long-term use of inhaled
corticosteroids and the risk of pneumonia in chronic obstruc-
tive pulmonary disease: a meta-analysis. Arch Intern Med
2009; 169:219 –229
37 Wedzicha JA, Calverley PMA, Seemungal TA, et al. The
prevention of chronic obstructive pulmonary disease exacer-
bations by salmeterol/fluticasone propionate or tiotropium
bromide. Am J Respir Crit Care Med 2008; 177:19 –26
38 Moher D, Cook DJ, Eastwood S, et al. Improving the quality
of reports of meta-analyses of randomized controlled trials:
the QUOROM statement. Lancet 1999; 354:1896 –1900
39 Sears M. Long-acting bronchodilators in COPD [editorial].
Chest 2008; 133:1057–1058
Original Research
 Safety and Efficacy of Combined Long-Acting β-Agonists and Inhaled
Corticosteroids vs Long-Acting β-Agonists Monotherapy for Stable
COPD : A Systematic Review
Gustavo J. Rodrigo, José A. Castro-Rodriguez and Vicente Plaza
Chest 2009;136; 1029-1038; Prepublished online July 24, 2009;
DOI 10.1378/chest.09-0821
This information is current as of July 14, 2010

Updated Information & Services

Updated Information and services can be found at:
http://chestjournal.chestpubs.org/content/136/4/1029.full.html
References
This article cites 31 articles, 20 of which can be accessed free at:
http://chestjournal.chestpubs.org/content/136/4/1029.full.html#ref-list-1
Citations Cited Bys
This article has been cited by 6 HighWire-hosted articles:
http://chestjournal.chestpubs.org/content/136/4/1029.f
ull.html#related-urls
Permissions & Licensing
Information about reproducing this article in parts (figures, tables) or in its entirety can be
found online at:
http://www.chestpubs.org/site/misc/reprints.xhtml
Reprints
Information about ordering reprints can be found online:
http://www.chestpubs.org/site/misc/reprints.xhtml
Citation Alerts
Receive free e-mail alerts when new articles cite this article. To sign up, select the
"Services" link to the right of the online article.
Images in PowerPoint format
Figures that appear in CHEST articles can be downloaded for teaching purposes in
PowerPoint slide format. See any online figure for directions.

Downloaded from chestjournal.chestpubs.org at Downey Community Hospital on July 14, 2010

© 2009 American College of Chest Physicians