Combined corticosteroid and long‐acting beta‐agonist in one inhaler versus placebo for chronic obstructive pulmonary disease
Abstract
Background
Long‐acting beta‐agonists and inhaled corticosteroids have both been recommended in guidelines for the treatment of chronic obstructive pulmonary disease. Their co‐administration in a combined inhaler may facilitate adherence to medication regimens, and improve efficacy.
Objectives
To assess the efficacy of combined inhaled corticosteroid and long‐acting beta‐agonist preparations, compared to placebo, in the treatment of adults with chronic obstructive pulmonary disease.
Search methods
We searched the Cochrane Airways Group Specialised Register of trials. The date of the most recent search is April 2007.
Selection criteria
Studies were included if they were randomised and double‐blind. Studies could compare any combined inhaled corticosteroids and long‐acting beta‐agonist preparation with placebo.
Data collection and analysis
Two authors independently assessed study risk of bias and extracted data. The primary outcomes were exacerbations, mortality and pneumonia. Health‐related quality of life (measured by validated scales), lung function and side‐effects were secondary outcomes. Dichotomous data were analysed as fixed effect odds ratios or rate ratios with 95% confidence intervals, and continuous data as mean differences and 95% confidence intervals.
Main results
Eleven studies met the inclusion criteria (6427 participants randomised). Two different combination preparations (fluticasone/salmeterol and budesonide/formoterol) were used. Study quality was good. Fluticasone/salmeterol and budesonide/formoterol both reduced the rate of exacerbations. Pooled analysis of both combination therapies indicated that exacerbations were less frequent when compared with placebo, Rate Ratio: 0.74 (95% CI 0.7 to 0.8). The clinical impact of this effect depends on the frequency of exacerbations experienced by patients. The patients included in these trials had on average 1‐2 exacerbations per year which means that treatment with combination therapy would lead to a reduction of one exacerbation every two to four years in these individuals. There is an overall reduction in mortality, but this outcome is dominated by the results of TORCH and further studies on budesonide/formoterol are required. The three year number needed to treat to prevent one extra death is 36 (95% CI 21 to 258), using a baseline risk of 15.2% from the placebo arm of TORCH. Both treatments led to statistically significant improvement in health status measurements, although the clinical importance of the differences observed is open to interpretation. Symptoms and lung function assessments favoured combination treatments. There was an increase in the risk of pneumonia with combined inhalers. The three year number needed to treat for one extra case of pneumonia is 13, using a baseline risk of 12.3% from the placebo arm of TORCH. Fewer participants withdrew from studies assessing combined inhalers due to adverse events and lack of efficacy.
Authors' conclusions
Compared with placebo, combination therapy led to a significant reduction of a quarter in exacerbation rates. There was a significant reduction in all‐cause mortality with the addition of data from the TORCH trial. The increased risk of pneumonia is a concern, and better reporting of this outcome in future studies would be helpful. In order to draw firmer conclusions about the effects of combination therapy in a single inhaler more data are necessary, particularly in relation to the profile of adverse events and benefits in relation to different doses of inhaled corticosteroids.
PICOs
Plain language summary
Combination therapy with inhaled corticosteroids and long‐acting beta‐agonists can reduce exacerbations and improve quality of life in people with chronic obstructive pulmonary disease (COPD) when compared to placebo treatment
Combinations of two classes of medication in one inhaler have been developed to treat people with COPD. Two types of combined inhaler exist currently: budesonide/formoterol (BDF ‐ 'Symbicort'), and fluticasone/salmeterol (FPS ‐ 'Advair' or 'Seretide'). The results of the studies showed that combined inhalers were effective and reduced the frequency of exacerbations compared with placebo medication to a level of three quarters of the previous rates. The patients included in these trials had on average 1‐2 exacerbations per year which means that treatment with combination therapy would lead to a reduction of one exacerbation every two to four years in these individuals. Combination therapy led to a reduction in mortality over three years, and also led to improvements in lung function and symptoms. However, there was an increased risk of pneumonia associated with combined inhalers, and further monitoring of this outcome in future trials would provide valuable information for consumers and clinicians. Future research is required to show whether there is a difference between combination inhalers with different strengths of inhaled corticosteroids.
Authors' conclusions
Background
Both inhaled corticosteroids and long‐acting beta‐agonists (LABAs) have been shown to be effective in chronic obstructive pulmonary disease (COPD). The evidence available indicates that there is a consistent, statistically significant difference in FEV1 of around 50 mL, and significant differences in health status measurements in favour of salmeterol in patients with a poor response to short‐acting bronchodilators (Appleton 2006). Inhaled steroids have not been shown to reduce the rate of decline in FEV1 although short‐term increases in FEV1 and significant reductions in exacerbations have been reported (Yang 2007). The impact of these effects on morbidity and mortality remain uncertain. LABAs are recommended by NICE 2004 and GOLD (GOLD 2006) when patients continue to experience problems on short‐acting bronchodilators.
The evidence base for the addition of long‐acting beta‐agonists to inhaled steroids in asthma is well established (Ni Chroinin 2004; Ni Chroinin 2005). There are several possible advantages to fixed combination therapy. The most obvious benefit is in terms of patient convenience with the expectation that this may lead to greater treatment adherence. Therefore, it is important to know whether inhaled corticosteroids combined with long‐acting beta‐agonist (LABAs) preparations reduce the number of exacerbations and improve other endpoints such as quality of life, despite the lack of effect on FEV1 decline.
This is an update of a previous review which considered the effect of combination therapy compared with placebo, and both monocomponents in people with chronic obstructive pulmonary disease (COPD) (Nannini 2004). The availability of several new studies has prompted us to split the review between comparisons with placebo and those with monocomponents. This review summarises the evidence from clinical trials comparing combination inhaled steroids and long‐acting beta‐agonists with placebo. Reviews of combination therapy with inhaled steroids (Nannini 2007a) and long‐acting beta‐agonists (Nannini 2007b) are published separately. Recent concerns have been raised regarding the safety of long‐acting beta‐agonists in asthma (Walters 2007) and also the validity of summary estimates from clinical trials which assess exacerbation rates without taking account of follow‐up time or adjustment for between‐patient variability (Suissa 2006).
Objectives
To determine the efficacy of combined inhaled corticosteroid and long‐acting beta‐agonists for stable COPD in comparison with placebo.
Methods
Criteria for considering studies for this review
Types of studies
Randomised, double‐blind clinical trials.
Types of participants
Adult patients (age > 45 years) with known, stable COPD fulfilling American Thoracic Society (ATS), European Respiratory Society (ERS) or Global Initiative for Chronic Obstructive Lung Disease (GOLD) diagnostic criteria. Patients were to be clinically stable, without evidence of an exacerbation for one month prior to study entry. Patients with significant diseases other than COPD, (for example with a diagnosis of asthma, cystic fibrosis, bronchiectasis, or other lung diseases) were excluded. However patients with partial reversibility on pulmonary function testing were included.
Types of interventions
-
Fluticasone/salmeterol versus placebo
-
Budesonide/formoterol versus placebo
Concomitant therapy was permitted, as long as there was no systematic difference between treatment groups.
Types of outcome measures
Primary outcomes
-
Exacerbations & hospitalisations
-
Mortality
-
Pneumonia has been added as a primary outcome for this update.
Secondary outcomes
-
Change in forced expiratory volume in 1 second (FEV1) and change in forced ventilatory capacity (FVC): trough, peak and average; and other measures of pulmonary function
-
Exercise performance ‐ six minute walk and other measures
-
Quality of life scales ‐ St George's Respiratory Questionnaire (SGRQ), Chronic Respiratory Disease Questionnaire (CRDQ)
-
Symptoms
-
Inhaled rescue medication used during the treatment period and other concomitant medication usage including antibiotics and steroids
-
Adverse events ‐ palpitations, tremor, hoarseness/dysphonia, oral candidiasis, cataracts, skin bruising, bone fracture, bone density, plasma cortisol level
Search methods for identification of studies
Electronic searches
Trials were identified using the Cochrane Airways Group Specialised Register of trials, which is derived from systematic searches of bibliographic databases including the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, CINAHL, AMED and PsycINFO, and handsearching of respiratory journals and meeting abstracts. All records in the Specialised Register coded as 'COPD' were searched using the following terms:
(((beta* and agonist*) and long*) or ((beta* and adrenergic*) and long*) and (*steroid or steroid* or corticosteroid*)) or (fluticasone and salmeterol) or Seretide or Advair or (formoterol and budesonide) or Symbicort
The most recent search on the Register was run in April 2007. In addition, we performed a search of LILACS (all years to March 2005) and CENTRAL (The Cochrane Library Issue 1, 2006).
Searching other resources
We reviewed reference lists of all primary studies and review articles for additional references. We also contacted authors of identified randomised trials about other published and unpublished studies. In addition, we consulted the online trial registries of GSK and AstraZeneca, manufacturers of FPS and BUDF respectively (www.ctr.gsk.co.uk; www.astrazenecaclinicaltrials.com).
Data collection and analysis
Selection of studies
Two reviewers independently identified abstracts of trials which appeared potentially relevant. Using the full text of each study, two reviewers independently selected trials for inclusion in the review. Agreement was by simple agreement; third party adjudication was used to resolve differences.
Data extraction and management
Two reviewers independently extracted data from included trials and entered results into the Cochrane Collaboration software program (Review Manager). In some cases, we estimated information regarding outcomes from graphs. This was performed independently by the two reviewers. Data extraction included the following items:
Population: Age, gender, smoking status, study setting (country, practice setting), inclusion and exclusion criteria.
Intervention: Dose, delivery device, duration
Control: Concurrent treatments (ipratropium, beta‐2 agonist, inhaled and systemic corticosteroids)
Outcomes: Pulmonary function measures (baseline and follow‐up FEV1 and FVC), timing of pulmonary function measures, 6‐minute walk, urgent visits, admissions, self‐rated symptom score/symptoms, quality‐of‐life instruments, adverse events (palpitations, dry mouth, blurred vision, urinary obstruction and constipation), assessors, adjudicator of clinical endpoints. Mortality outcome data were collected from studies of greater than one year's duration where these were available.
Design: method of randomisation, presence and type of run‐in period, study design (parallel, cross‐over)
Assessment of risk of bias in included studies
We assessed the risk of bias for each study according to the extent to which the design of the trial had protected the findings from bias. Our assessments we made on the basis of allocation, its concealment and blinding.
Dealing with missing data
The reported confidence interval or P value were used to calculate standard deviations, or standard errors, for results when these were not reported and could not be obtained from the authors of the papers.
Assessment of heterogeneity
For pooled effects, heterogeneity was to be tested using I2 measurement of the degree of variation between the studies, not attributable by the play of chance.
Data synthesis
For continuous variables, a fixed effects weighted mean difference (MD) was used for outcomes measured on the same metric. Standardised mean difference (SMD) and 95% confidence interval (CI) was calculated for outcomes where data were combined from studies using different metrics. All similar studies were to be pooled using fixed effect MD/SMD and 95% CIs.
For dichotomous variables, a fixed effect odds ratio (OR) with 95% confidence intervals (95% CI) were calculated for individual studies. All similar studies were pooled using fixed effect OR and 95% CIs. Where mean treatment differences were reported, data were entered as generic inverse variance (GIV), provided a standard error for the difference could be extracted or imputed. Where this method was used the effect size was reported from the original papers, for example as Rate Ratio. This method (GIV) was not available when the protocol was written for the review so was not pre‐specified.
Subgroup analysis and investigation of heterogeneity
Whilst we separated the type of steroid and long‐acting beta‐agonist, we pooled studies with differing dosages of the same drug. We planned a priori subgroups as:
-
Disease severity (related to baseline FEV1 and placebo group exacerbation rate) according to the GOLD staging = IIA, IIB (moderate COPD, characterised by deteriorating lung function (A = FEV1</=80% predicted; B = </=50% predicted) and progression of symptoms) and III (severe COPD, characterised by severe airflow limitation (FEV1<30% predicted) and presence of respiratory failure or clinical signs of right heart failure (GOLD 2006).
-
Prior inhaled corticosteroid plus long‐acting beta agonists use (dichotomised as yes/no).
-
Concurrent therapy with routine beta‐agonist use (short or long‐acting), corticosteroid (systemic or inhaled) or theophylline use (dichotomised as yes/no).
-
Reversibility of airflow obstruction with beta‐2‐agonist therapy (dichotomised as partial/none). Definition: >12% and > 200 ml from baseline FEV1 or > 12% as a per cent of the predicted normal value following MDI salbutamol 200 to 400
-
Dose, duration and delivery method of therapy.
Sensitivity analysis
For pooled effects, heterogeneity was to be tested using I2 measurement of the degree of variation between the studies, not attributable by the play of chance. If heterogeneity was found, (I2 statistic more than 20%), a random effects model was also used to determine the impact of heterogeneity on the overall pooled effect. In addition, the robustness of the results was tested using a sensitivity analysis based on the quality of the trials where possible.
Results
Description of studies
Results of the search
For details of search history, see Table 1. An overview of the division of our original review into three and description of the search results is given in Figure 1.
Flow chart to illustrate separation of review between three comparisons. Six RCTs met the original entry criteria of the review. All of these had a placebo and long‐acting beta‐agonist arm, and five assessed combination against steroids. Seven new studies with one or more control comparisons were identified: five had a placebo arm, three had a long‐acting beta‐agonist arm, and two had an inhaled steroid treatment arm.
| Version | Detail |
| 1st published version ‐ Issue 4, 2003 (All years to April 2002) | References identified: 34 |
| 2nd published version ‐ Issue 3, 2004 (April 2003‐April 2004) | References identified: 12 |
| 3rd published version ‐ Issue 3, 2005 (April 2004‐April 2005) | References identified: 52 |
| 4th published version ‐ April 2005 ‐ April 2007 | References identified: 66 |
Included studies
Eleven studies are included in this review. One study is ongoing (Morgan 2004). For a full description of baseline characteristics, methods used and inclusion and exclusion entry criteria. individual studies, see Characteristics of included studies.
Design
All trials had a randomised, double‐blind parallel group design. Methods of randomisation were described in one study (Mahler 2002). Method of blinding was not fully described in all studies. Following correspondence from GSK, trial methodology was confirmed for TRISTAN, and AstraZeneca confirmed methodology for Szafranski 2003. Study characteristics were sufficiently described in three data sets without journal publication to justify their inclusion in the review (O'Donnell 2006; SFCT01; SCO100540).
Participants
A total of 6427 participants were randomised to studies in the review. Participants suffered from COPD, with variable definition of COPD and reversibility. COPD was defined by national or international criteria: ATS (Hanania 2003; Mahler 2002) ERS (TORCH; TRISTAN) and GOLD (Barnes 2006; Calverley 2003; Dal Negro 2003; SCO100540; Szafranski 2003). In two studies, definitions were based on lung function tests and smoking history (O'Donnell 2006; SFCT01). Patient populations in the studies suffered from moderate and severe COPD. Hanania 2003 and Mahler 2002 enrolled participants with reversible and non‐reversible COPD. In TORCH the participants were not required to have had previous exacerbations requiring oral steroids or antibiotics to be included in the study.
Interventions
All eleven studies compared combination therapy with placebo.
In one study all participants had a two‐week run‐in treatment with oral corticosteroids, inhaled formoterol and prn SABA (Calverley 2003).
Concomitant therapy was as‐needed short‐acting beta agonist, or oral steroids and/or antibiotics in the case of exacerbations. However in three studies, theophylline was also used. Eleven percent of participants in Hanania 2003, and all 18 participants in Dal Negro 2003 received theophylline, in addition to the study drugs. The exact proportion of patients in TRISTAN who were taking theophylline was not reported.
In two studies the combination of inhaled corticosteroid/long‐acting beta‐agonist was fluticasone/salmeterol (FPS) was 250 mcg/50 mcg twice daily (Dal Negro 2003; Hanania 2003). In the remainder of the FPS studies the dose was 500 mcg/50 mcg twice daily. In Calverley 2003 and Szafranski 2003 the combination inhaled corticosteroid/long‐acting beta agonist was budesonide/formoterol (BDF) (320 mcg/9 mcg twice daily).
Duration
13 weeks or less: Barnes 2006; O'Donnell 2006.
24 weeks: Hanania 2003; Mahler 2002; SCO100540.
52 weeks: Calverley 2003; Dal Negro 2003; Szafranski 2003; SFCT01; TRISTAN.
156 weeks: TORCH.
Outcomes
Exacerbations were stratified by medication given (oral steroid and/or antibiotic treatment in Barnes 2006; Calverley 2003; Dal Negro 2003; SCO100540; SFCT01; Szafranski 2003; TORCH; TRISTAN) or hospitalisation (Barnes 2006; Dal Negro 2003; SCO100540; TORCH; TRISTAN). O'Donnell 2006 did not record/report exacerbation data. Hanania 2003 and Mahler 2002 withdrew participants who exacerbated. Lung function was measured as FEV1 or PEF in all the studies. Quality of life assessment by the SGRQ or CRDQ were available for Calverley 2003; Hanania 2003; Kardos 2007; Mahler 2002; SCO100540; SFCT01; Szafranski 2003; TORCH; TRISTAN. All cause mortality was reported by TORCH.
Excluded studies
Studies that did not meet the entry criteria of this review are listed in Characteristics of excluded studies.
Risk of bias in included studies
Our assessments of the risk of bias for allocation and blinding are provided in Figure 2. The majority of our judgements on allocation procedures were unclear due to paucity of this information in the trial reports. We were able to ascertain that these were at a low risk of bias in three large studies (Szafranski 2003; TORCH; TRISTAN). All studies used identical inhaler devices to deliver therapy in the treatment groups.
Methodological quality summary: review authors' judgements about each methodological quality item for each included study.
Effects of interventions
See Summary of Findings table (See Figure 3).
Summary of findings table for exacerbations, mortality, quality of life, pneumonia and adverse events
Primary outcomes: Exacerbations, Mortality and Pneumonia
Rate of exacerbations
FPS versus placebo
There was a significant reduction in the rate of exacerbations with combination therapy when compared with placebo (0.74; 95% CI 0.69 to 0.8, three trials, N = 4222). This result was not altered by removing TRISTAN whose summary estimate may have been biased by inadequate adjustment for between patient variability (Suissa 2006), see Figure 4. Additional analysis was performed on exacerbations by specific definitions. Compared with placebo, FPS led to fewer exacerbations requiring oral steroids (rate ratio 0.57; 95% CI 0.52 to 0.63, three studies), less requirement for antibiotics (rate ratio 0.6 95% CI 0.43 to 0.84, one study) and fewer hospitalisations (rate ratio 0.83; 95% CI 0.7 to 0.97, two studies).
Forest plot of pooled rate ratios for combination fluticasone and salmeterol and placebo, with the summary estimate for TRISTAN included (primary analysis) and removed (sensitivity analysis)
BDF versus placebo
There was a significant effect on pooled exacerbation rates in favour of BDF when expressed as a rate ratio compared with placebo (0.74; 95% CI 0.62 to 0.88) (see Figure 4). These results are based on 913 participants from two trials (Calverley 2003; Szafranski 2003).
When combined together the overall reduction in rate ratio of exacerbations using either FPS or BDF is 0.74; 95% CI 0.69 to 0.79, five trials, N = 5135.
Dichotomous data (number of people experiencing an exacerbation)
FPS versus placebo
No significant difference: OR 0.91; 95% CI 0.69 to 1.19, four studies, N = 1109.
Mortality
FPS versus placebo
The adjusted hazard ratio from TORCH did not identify a significant effect of FPS over placebo (0.825, 0.681 to 1.002, P = 0.052). Analysed as number of deaths in each treatment group by odds ratio and combined with data from other four other studies, there was a significant reduction in the odds of death in favour of FPS versus placebo (Odds Ratio 0.79, 95% CI 0.65 to 0.98, N = 4829). Data were separated according to the time point and subgrouped for data at three years, data reported at 1‐3 years, data at 1 year and data at 6 months.
BDF versus placebo
The two studies of one year duration did not detect a significant difference between BDF and placebo. The point estimate (Odds Ratio 0.78, 95%CI 0.35 to 1.73) was similar to that for FPS and the was no significant difference in the odds ratio for mortality between FPS and BDF.
When combined together the overall reduction in the odds ratio for mortality with either FPS or BDF was 0.79, 95% CI 0.65 to 0.96, N = 5752, see Figure 5. Since differing lengths of follow‐up across the studies hinders the calculation of a pooled NNT, we have tabulated this for each study individually (see Table 2). The three year NNT (using the baseline risk of 15.2% in the placebo arm of TORCH) to prevent one extra death is 36 (95% CI 21 to 258).
Pooled effect estimate on mortality for all combined inhalers versus placebo
| Study ID | Study duration | Placebo rate (%) | NNT |
| 156 weeks | 15.2 | 42 (24 to 387) | |
| 24 weeks | 0 | NA | |
| 52 weeks | 0 | NA | |
| 52 weeks | 1.94 | 292 (169 to 2628) | |
| 24 weeks | 1.66 | 293 (176 to 1540) | |
| 52 weeks | 1.95 | 249 (149 to 1307) | |
| 52 weeks | 4.5 | 110 (66 to 581) |
Pneumonia
FPS versus placebo
Addition of data from TORCH shows a significant increase in pneumonia with FPS in comparison with placebo. The pooled OR is 1.80 (95% CI 1.48 to 2.18).
BDF versus placebo
Calverley 2003 reported data on pneumonia and although the increase was not statistically significant (OR 4.13, 95% CI 0.87 to 19.64) the direction of effect was similar to FPS and there is no significant difference between the increase in pneumonia when BDF is compared to FPS.
For combined inhalers, the pooled OR for pneumonia is 1.83 (95% CI 1.51 to 2.21). Table 3 gives the range of NNT(H) across the studies for pneumonia. A pooled NNT(H) was not calculated due to the wide differences in duration and the likely impact this would have on the calculation of a pooled event rate. The three year NNT(H) (using the baseline risk of 12.3% in the combination therapy arm of TORCH) for one extra patient to suffer pneumonia was 13.
| Study ID | Study duration | Placebo rate (%) | NNT(h) |
| 24 weeks | 0 | NA | |
| 24 weeks | 0 | NA | |
| 13 weeks | 0 | NA | |
| 24 weeks | 0 | NA | |
| 8 weeks | 1.56 | 83 | |
| 156 weeks | 12.3 | 13 | |
| 52 weeks | 0.83 | 153 | |
| 52 weeks | 3.6 | 36 |
Secondary outcomes
Quality of life
St George Respiratory Questionnaire (results collected from end of study)
FPS versus placebo
Treatment with FPS improved SGRQ scores by an average of ‐2.9 units versus placebo (95% CI: ‐3.61 to ‐2.18, four studies, N = 3346).
BDF versus placebo
There was a significant effect in favour of BDF compared with placebo of ‐6.06 units on the SGRQ (95%CI ‐7.90 to ‐4.22). There was a high level of heterogeneity when these data were pooled (I square 71.8%). Random effects modelling also generated a significant effect (MD ‐5.8 units [95%CI ‐9.32 to ‐2.28]). The magnitude of improvement in the Szafranski 2003 BDF group was 3.9 units from baseline, and was not dissimilar to the change scores from post run‐in treatment in the Calverley 2003 (see graphical presentation of data in the published article, P 916). However, the placebo group deteriorated more in Calverley 2003, which possibly reflects the withdrawal of active treatment, with the subsequent loss of the pre‐dosing effects achieved with high dose OCS and LABAs. In comparison, BDF may have maintained the pre‐dosing treatment effects of QoL more successfully.
Chronic Respiratory Disease Questionnaire (results collected from end of study)
FPS versus placebo
Pooled data from Mahler 2002 and Hanania 2003 indicated a statistically significant improvement in CRDQ for those treated with FPS compared with placebo (5 units (95% CI 2.48 to 7.52).
Symptom score
FPS versus placebo
FPS led to improved symptom scores (transitional dysponea index) when compared with placebo (MD 1.04 (95% CI 0.56 to 1.53) Mahler 2002 and Hanania 2003 gave a significant difference in change scores in favour FPS .
TRISTAN reported improvements in symptoms after treatment in favour of FPS versus placebo on breathlessness scores (P < 0.001), and night time awakenings (FPS mean number of nights per week: 2.31; placebo mean: 3.01 (P = 0.006). Cough scores also favoured FPS (P = 0.018).
BDF versus placebo
Data were pooled for Calverley 2003 and Szafranski 2003. There was a significant effect in favour of BDF when compared with placebo, MD ‐0.63 (95%CI ‐0.90 to ‐0.37).
Lung function
Predose & post‐dose FEV1 ‐ Change from baseline
FP versus placebo
Pooled analysis of data were conducted without findings from the Dal Negro 2003 study. Owing to the small size of this study, we were concerned that the SDs represented an inaccurate estimate for the SDs of the population, and the small variance increased the weight of the study, out of all proportion to its size. Data pooled from seven studies gave a MD in predose FEV1 of 0.16 Litres; 95% CI 0.14 to 0.19, N = 1408. Analysis of post‐dose FEV1 from TORCH indicated a significant improvement in favour of FPS over placebo of 0.09 L.
BDF versus placebo
FEV1 data were reported as % change. There was a significant increase in FEV1 in favour of BDF versus placebo of 14.40% (95%CI 11.91 to 16.90, two studies).
Predose FEV1‐ Absolute values
FPS versus placebo
TRISTAN and SFCT01 reported data as mL and Dal Negro 2003 reported % predicted.
The results of this analysis have not been pooled due to concern over the small number of similar patients in Dal Negro 2003, and the very low variance. It is therefore not appropriate to combine the results on an SMD scale.
Pooled data from TRISTAN and SFCT01 led to a significant difference in favour of FPS of 0.13 Litres; 95% CI 0.10 to 0.16.
Rescue medication
FPS versus placebo
Pooled data from Mahler 2002 and Hanania 2003 indicated a significant reduction in mean puffs per day of short‐acting beta‐agonist usage for FPS versus placebo (MD ‐1.19 puffs/day; 95% CI: ‐1.83 to ‐0.55).
Mahler 2002 reported significant increases in the percentage of nights with no awakenings requiring short‐acting beta‐agonist in favour of FPS versus placebo (5.7% versus ‐4.3% respectively, P < 0.031).
TRISTAN reported a significant difference in median % of days without use of relief medication: FPS: 14% versus placebo: 0% (P < 0.001).
BDF versus placebo
BDF treatment reduced the requirement for reliever medication when compared with placebo. The combined results of Calverley 2003; Szafranski 2003 , MD ‐0.87 puffs per day (95%CI ‐1.15 to ‐0.58), I2 29.1%; Random effects modelling: MD ‐0.92 puffs per day (95% CI ‐1.34 to ‐0.50).
Safety & tolerability
FPS versus placebo
There was no significant difference in the occurrence of overall reported adverse events between FPS and placebo: OR 1.10; 95% CI 0.96 to 1.27; eight studies, N = 5493.
Pneumonia, candidiasis, nasopharyngitis, hoarseness and upper respiratory tract infections (URTI) occurred more frequently in FPS treated participants:
Pneumonia: OR 1.80; 95% CI 1.48 to 2.18; seven studies, N = 5229.
Candidiasis: OR 5.73; 95%CI 3.07 to 10.67; six studies, N = 1958.
Hoarseness: OR 8.79; 95% CI 1.11 to 69.62; two studies, N = 585.
Nasopharyngitis: OR 1.28; 95% CI 1.05 to 1.56; two studies, N = 3535.
URTI: OR 1.23; 95% CI 1.04 to 1.47; five studies, N = 4963.
Withdrawals
FPS versus placebo
There were significantly fewer withdrawals from treatment with FPS than placebo (OR 0.69; 95% confidence interval 0.62 to 0.78). Withdrawals due to adverse events and lack of efficacy also occurred less frequently on treatment with FPS than placebo (withdrawal due to adverse event: OR 0.76; 95% confidence interval 0.65 to 0.88; eight studies, 5110 participants; withdrawal due to lack of efficacy: OR 0.30; 95% confidence interval 0.21 to 0.42; five studies, 4620 participants).
BDF versus placebo
Data were pooled from Calverley 2003 and Szafranski 2003 for withdrawals due to worsening COPD symptoms and adverse events.
There was a significant difference in withdrawals due to worsening of COPD symptoms when BDF was compared with placebo, OR 0.40 [95%CI 0.28 to 0.58].
There was no significant difference between BDF and placebo on the likelihood withdrawal due to adverse events other than COPD deterioration, OR 1.30 [95%CI 0.78 to 2.18].
Discussion
We have reviewed data from 11 randomised controlled trials (6427 participants) assessing the effectiveness of combined inhaled corticosteroid and long‐acting beta‐agonist in the treatment of chronic obstructive pulmonary disease (COPD). Whilst the consensus regarding the definitions of COPD and COPD exacerbations evolve, the trial evidence to date indicates that in severe COPD where this is defined by low FEV1, significant smoking history, and a recent history of reduced disease control there is a role for combination therapy in reducing exacerbations which lead to unscheduled additional treatment, including courses of oral steroid therapy and hospitalisation.
Patient populations
There is much debate as whether 'chronic obstructive pulmonary disease' includes patients who have a significant bronchodilator response to short‐acting beta‐agonists, although recent definitions permit some reversibility. In this review, we have attempted to stratify data according to baseline reversibility. There is debate as to whether such a distinction is a clinically valid dichotomy (Burge 2003; Calverley 2003a). We have opted to pool data from both these populations as there remains little consensus on this issue (ATS 1995), but for some outcomes (such as FEV1) a sub‐group analysis has been carried out on the basis of reversibility. It is also of interest that Calverley 2003 and Szafranski 2003 were classified as low reversibility on the basis of change in FEV1 following short acting beta‐agonists, but nevertheless achieved a mean change in FEV1 of around 15% following pre‐treatment with oral steroids and long‐acting beta‐agonist treatment in the studies.
Baseline FEV1 data from the studies indicate that the effects observed in the analyses were measured in populations who were overwhelmingly severe. Intention‐to‐treat analyses were conducted in all the studies, but on outcomes such as mean exacerbation rates, the withdrawal of severe frequent exacerbators in the studies may have distorted study findings due to the lower exacerbation rates on active treatment. The loss of these patients from the studies may thus limit the accuracy of mean event rates.
There was also a high proportion of current smokers in Mahler 2002 and TRISTAN. A recent study has demonstrated a diminished response to dexamethasone in vitro following the simulation of smoking effects. Smoke effects induced the inhibition of the histone deacetylases (HDAC) enzyme which is necessary in the process of gene repression (Ito 2001). Whilst one in vivo study has demonstrated the resistance to steroid action in asthmatic smokers (Pedersen 1996), there may nevertheless be a similar inhibition of the expression of corticosteroids when inhaled or systemic steroids are used in current smokers with COPD, resulting in compromised corticosteroid effectiveness.
Clinical endpoints
The outcomes identified a priori to be of primary significance related to exacerbations and hospitalisations. Limited data were presented in the trials on hospitalisations as a separate outcome although the evidence available did indicate that combination treatment reduced the rate of this event. Previous studies have indicated that ICS reduces the mean rate of exacerbations requiring oral steroid or antibiotics in patients with moderate and severe COPD (Yang 2007). However, the dose at which ICS is considered to efficacious, and the patients most likely to respond to this therapy remain unclear. The magnitude of these differences and their clinical impact depends upon the frequency of patients exacerbations. In severe patients with a history of one or more exacerbations in the previous year, the difference of 24% may translate to one exacerbation less every 4 years. Further assessment of the impact of the combination therapy upon hospitalisations and exacerbations is required.
There is growing interest in the definition and analysis of exacerbations which occur in clinical trials of therapies in COPD (Roisin 2000; Suissa 2006; Calverley 2005). The role of any maintenance therapy in this population should be assessed in terms of the burden of deteriorating disease control to both the individual patient and the healthcare system which rationalises resources in treating the debilitating effects of chronic disease progression. The relationship between event defined exacerbations (such as short courses of oral steroids or antibiotics, unscheduled presentation) and other markers of health such as symptoms and quality of life continue to present trialists and guideline developers with the dilemma of when to recommend initiating additional therapy based on the trial evidence available, and the extent to which varying definitions of endpoint alter external validity in applying evidence to the needs of an individual patient. The question of exacerbations, and the appropriate statistical analysis of rate ratios, casts some doubt over the validity of some of the findings in this review. In particular, the large long‐term studies (i.e. those in excess of six months) which are adequately powered to detect statistically significant findings may overestimate the treatment effects of this therapy (Suissa 2006). The method of weighting counts of exacerbations described by Suissa 2006 (using duration of person follow‐up time as a denominator in calculating the mean group rate of exacerbations rather than an unweighted approach) was undertaken in Calverley 2003; Szafranski 2003 and TORCH. In these studies the effects were consistent and significantly favoured combination therapy over placebo.
Mortality was recorded in many of the studies, but was collected and analysed as a primary outcome in TORCH. Cause‐specific mortality was reported, but the definition of a primary cause of death continues to pose challenges in a population of patients who suffer from comorbidities such as lung cancer (McGarvey 2007). The findings in the original trial publication as a hazard ratio (HR). The HR reported in TORCH was adjusted for repeated measurements, and was not statistically significant (ratio: 0.825, P = 0.052). The significant finding in our review of a ratio of odds of death overall of 0.79 in favour of treatment probably reflects the fact that our analyses were not subject to adjustment for repeated measurement. Furthermore, this estimate encompasses data from studies with considerably shorter lengths of follow‐up which provide only limited long‐term prognostic value. This explains why TORCH provides 53% participants, and 89% of the weight. The three year NNT (from the baseline risk in TORCH) to prevent one extra death is 36 (95% CI 21 to 258).
Both the CRDQ and the SGRQ were used to measure the effect of combination therapy on health status. There was a high level of statistical variation in the FPS studies for total SGRQ scores, although the reasons for this are not clear. Differing treatment protocols in the two budesonide/formoterol combination studies may explain the differences between health status scores. Although differences between combination therapy and placebo in Calverley 2003 and Szafranski 2003 were statistically significant, the difference from Calverley 2003 suggested that combination treatment led to a clinically meaningful advantage compared with placebo (i.e. greater than 4 units, Jones 2002). This difference should be interpreted with some caution due to the effects of pre‐treatment with high dose prednisolone and formoterol, which led to a clinically meaningful increase of 4.5 units on the SGRQ. These effects were sustained with combination treatment when compared with placebo, but statistical tests on the comparisons with component therapy were not available.
Side‐effects were more common on treatment than placebo. The increased risk of pneumonia with combination therapy is of particular concern since this frequently leads to hospitalisation in elderly patients. The association between prescription of inhaled fluticasone and pneumonia in people with COPD has been identified from a case‐control study (Ernst 2007). The rate ratio from that study indicated an increase in risk of pneumonia leading to hospitalisation of 70% (RR 1.7). From RCT data in this review the combination of salmeterol and fluticasone also led to an increase in the odds of pneumonia of 1.8 (representing an increase in the risk of pneumonia of 65%). The NNT(h) for this outcome varies considerably, with three long‐term studies (i.e. one year or greater) showing the lowest NNT(H) (Calverley 2003: 36; TORCH: 13; TRISTAN: 153, see Table 3).
The advantages associated with combination therapy extend beyond potentially greater patient compliance with a convenient delivery system of two therapies. The component treatments have complementary mechanisms of action, whereby the smooth muscle relaxation of the beta‐agonist combined with the anti‐inflammatory action of the steroid offer greater scope of clinical benefits than they would do alone. Cazzola 2000 reported a single‐blind trial assessing the additive benefit of long‐acting beta‐agonist and inhaled steroids. When administered in separate inhalers there was no significant improvement in lung function until three months of treatment. The results from the trials in this review have not addressed the question of synergistic effects, and further research comparing combination therapy in a single versus separate inhaler devices could elucidate this issue more clearly.
Fixed dose therapy may be an effective pharmacological strategy in clinically stable moderately severe COPD, but if patients become unstable other approaches would be needed (GOLD 2006). Consensus guidelines recommend oral corticosteroids and/or antibiotics depending on the cardinal symptoms of the exacerbation of COPD (GOLD 2006). At this time it would be inappropriate to manage a clinically unstable situation with combined fixed therapy on its own.
Flow chart to illustrate separation of review between three comparisons. Six RCTs met the original entry criteria of the review. All of these had a placebo and long‐acting beta‐agonist arm, and five assessed combination against steroids. Seven new studies with one or more control comparisons were identified: five had a placebo arm, three had a long‐acting beta‐agonist arm, and two had an inhaled steroid treatment arm.
Methodological quality summary: review authors' judgements about each methodological quality item for each included study.
Summary of findings table for exacerbations, mortality, quality of life, pneumonia and adverse events
Forest plot of pooled rate ratios for combination fluticasone and salmeterol and placebo, with the summary estimate for TRISTAN included (primary analysis) and removed (sensitivity analysis)
Pooled effect estimate on mortality for all combined inhalers versus placebo
Comparison 1 Combined inhalers versus Placebo (Primary Outcomes), Outcome 1 Exacerbation Rates with combined inhalers v. placebo.
Comparison 1 Combined inhalers versus Placebo (Primary Outcomes), Outcome 2 Mortality.
Comparison 1 Combined inhalers versus Placebo (Primary Outcomes), Outcome 3 Pneumonia.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 1 Number of participants with one or more exacerbation.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 2 End of treatment mean number of exacerbations per participant.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 3 Exacerbations.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 4 Number of exacerbations by type.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 5 Exacerbations by type.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 6 Mortality.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 7 Change from baseline in St George's Respiratory Questionnaire (total score).
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 8 Change from baseline in St George's Respiratory Questionnaire (domain ‐ symptoms).
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 9 Change from baseline in St George's Respiratory Questionnaire (domain ‐ activity).
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 10 Change from baseline in St George's Respiratory Questionnaire (domain ‐ impact).
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 11 End of treatment St George's Respiratory Questionnaire scores (total score).
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 12 End of treatment St George's Respiratory Questionnaire scores (domain ‐ breathlessness).
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 13 Change from baseline in Canadian Respiratory Disease Questionnaire scores.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 14 Change from baseline in Transitional Dyspnoea Index (TDI) scores.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 15 End of treatment symptom scores.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 16 Change from baseline in predose FEV1.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 17 Change from baseline in postdose FEV1.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 18 End of treatment FEV1 (% predicted).
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 19 End of treatment FEV1 (Litres).
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 20 End of treatment am PEF (L/min).
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 21 End of treatment shuttle walk test.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 22 End of treatment rescue medication usage (puffs/day).
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 23 Change from baseline in rescue medication usage (puffs/day).
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 24 Adverse events ‐ any event.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 25 Adverse events ‐ candidiasis.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 26 Withdrawals due to adverse events.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 27 Withdrawals due to lack of efficacy/exacerbations.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 28 Withdrawals.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 29 Adverse events ‐ pneumonia.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 30 Adverse events ‐ nasopharyngitis.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 31 Adverse events ‐ pharyngolaryngeal pain.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 32 Adverse events ‐ upper respiratory tract infection.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 33 Adverse events ‐ headache.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 34 Adverse events ‐ hoarseness.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 35 Adverse events ‐ pyrexia.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 36 Adverse events ‐ cough.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 37 Adverse events ‐ palpitations.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 38 Adverse events ‐ mouth ulceration.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 39 Adverse events ‐ toothache.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 40 Adverse events ‐ urinary tract infection.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 41 Adverse events ‐ dyspnoea.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 42 Adverse events ‐ blood glucose increased.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 43 Adverse events ‐ insomnia.
Comparison 2 Fluticasone/salmeterol (FPS) versus placebo (PLA), Outcome 44 Adverse events ‐ bronchitis.
Comparison 3 Budesonide/formoterol (BDF) versus placebo (PLA), Outcome 1 Severe Exacerbations.
Comparison 3 Budesonide/formoterol (BDF) versus placebo (PLA), Outcome 2 Mean severe exacerbation rates per patient per year.
Comparison 3 Budesonide/formoterol (BDF) versus placebo (PLA), Outcome 3 Quality of life ‐ SGRQ (change scores).
Comparison 3 Budesonide/formoterol (BDF) versus placebo (PLA), Outcome 4 Quality of life ‐ change scores.
Comparison 3 Budesonide/formoterol (BDF) versus placebo (PLA), Outcome 5 Rescue medication usage.
Comparison 3 Budesonide/formoterol (BDF) versus placebo (PLA), Outcome 6 Symptoms (change scores).
Comparison 3 Budesonide/formoterol (BDF) versus placebo (PLA), Outcome 7 Mean FEV1 (% change from baseline).
Comparison 3 Budesonide/formoterol (BDF) versus placebo (PLA), Outcome 8 Adverse events ‐ 'serious events'.
Comparison 3 Budesonide/formoterol (BDF) versus placebo (PLA), Outcome 10 Adverse events ‐ pneumonia.
Comparison 3 Budesonide/formoterol (BDF) versus placebo (PLA), Outcome 11 Withdrawals due to worsening COPD symptoms.
Comparison 3 Budesonide/formoterol (BDF) versus placebo (PLA), Outcome 12 Withdrawals due to adverse events.
Comparison 3 Budesonide/formoterol (BDF) versus placebo (PLA), Outcome 13 Mortality.
| Version | Detail |
| 1st published version ‐ Issue 4, 2003 (All years to April 2002) | References identified: 34 |
| 2nd published version ‐ Issue 3, 2004 (April 2003‐April 2004) | References identified: 12 |
| 3rd published version ‐ Issue 3, 2005 (April 2004‐April 2005) | References identified: 52 |
| 4th published version ‐ April 2005 ‐ April 2007 | References identified: 66 |
| Study ID | Study duration | Placebo rate (%) | NNT |
| 156 weeks | 15.2 | 42 (24 to 387) | |
| 24 weeks | 0 | NA | |
| 52 weeks | 0 | NA | |
| 52 weeks | 1.94 | 292 (169 to 2628) | |
| 24 weeks | 1.66 | 293 (176 to 1540) | |
| 52 weeks | 1.95 | 249 (149 to 1307) | |
| 52 weeks | 4.5 | 110 (66 to 581) |
| Study ID | Study duration | Placebo rate (%) | NNT(h) |
| 24 weeks | 0 | NA | |
| 24 weeks | 0 | NA | |
| 13 weeks | 0 | NA | |
| 24 weeks | 0 | NA | |
| 8 weeks | 1.56 | 83 | |
| 156 weeks | 12.3 | 13 | |
| 52 weeks | 0.83 | 153 | |
| 52 weeks | 3.6 | 36 |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Exacerbation Rates with combined inhalers v. placebo Show forest plot | 5 | Rate Ratio (Fixed, 95% CI) | 0.74 [0.69, 0.79] | |
| 1.1 Fluticasone/salmeterol | 3 | Rate Ratio (Fixed, 95% CI) | 0.74 [0.69, 0.80] | |
| 1.2 Budesonide/formoterol | 2 | Rate Ratio (Fixed, 95% CI) | 0.74 [0.62, 0.88] | |
| 2 Mortality Show forest plot | 7 | 5752 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.79 [0.65, 0.96] |
| 2.1 Fluticasone/salmeterol | 5 | 4829 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.79 [0.65, 0.97] |
| 2.2 Budesonide/formoterol | 2 | 923 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.78 [0.35, 1.73] |
| 3 Pneumonia Show forest plot | 8 | 5739 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.83 [1.51, 2.21] |
| 3.1 FPS | 7 | 5229 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.80 [1.48, 2.18] |
| 3.2 BDF | 1 | 510 | Odds Ratio (M‐H, Fixed, 95% CI) | 4.13 [0.87, 19.64] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Number of participants with one or more exacerbation Show forest plot | 5 | 1235 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.88 [0.67, 1.15] |
| 1.1 Reversible population | 1 | 126 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.32 [0.06, 1.66] |
| 1.2 Partially reversible population (mixed population) | 2 | 713 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.00 [0.69, 1.44] |
| 1.3 Poorly reversible population | 2 | 396 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.80 [0.53, 1.21] |
| 2 End of treatment mean number of exacerbations per participant Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 2.1 Partially reversible population (mixed population) | 0 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 Poorly reversible population | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3 Exacerbations Show forest plot | 3 | Rate ratio (Fixed, 95% CI) | 0.74 [0.69, 0.80] | |
| 3.1 Partially reversible population (mixed population) | 0 | Rate ratio (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.2 Poorly reversible population | 3 | Rate ratio (Fixed, 95% CI) | 0.74 [0.69, 0.80] | |
| 4 Number of exacerbations by type Show forest plot | 2 | Odds Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 4.1 Requirement for oral steroids | 2 | 417 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.01 [0.61, 1.68] |
| 4.2 Requirement for antibiotic treatment | 1 | 140 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.80 [0.26, 2.44] |
| 4.3 Requirement for oral steroid or antibiotic treatment | 1 | 140 | Odds Ratio (M‐H, Fixed, 95% CI) | 3.32 [0.13, 82.80] |
| 4.4 Hospitalisation | 1 | 140 | Odds Ratio (M‐H, Fixed, 95% CI) | 3.32 [0.13, 82.80] |
| 5 Exacerbations by type Show forest plot | 3 | Rate ratio (Fixed, 95% CI) | Subtotals only | |
| 5.1 Requirement for oral steroids | 3 | Rate ratio (Fixed, 95% CI) | 0.57 [0.52, 0.63] | |
| 5.2 Requirement for antibiotic treatment | 1 | Rate ratio (Fixed, 95% CI) | 0.60 [0.43, 0.84] | |
| 5.3 Requirement for oral steroid or antibiotic treatment | 0 | Rate ratio (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5.4 Hospitalisation | 2 | Rate ratio (Fixed, 95% CI) | 0.83 [0.70, 0.97] | |
| 6 Mortality Show forest plot | 5 | 4829 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.79 [0.65, 0.97] |
| 6.1 Mortality: three year data | 1 | 3057 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.81 [0.66, 0.99] |
| 6.2 Mortality: >one and <three year data | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 6.3 Mortality: one year data | 3 | 1426 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.63 [0.21, 1.90] |
| 6.4 Mortality: 6 month data | 1 | 346 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.15 [0.01, 3.01] |
| 7 Change from baseline in St George's Respiratory Questionnaire (total score) Show forest plot | 4 | SGRQ units (Fixed, 95% CI) | ‐2.90 [‐3.61, ‐2.18] | |
| 7.1 Partially reversible population (mixed population) | 0 | SGRQ units (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 7.2 Poorly reversible population | 4 | SGRQ units (Fixed, 95% CI) | ‐2.90 [‐3.61, ‐2.18] | |
| 8 Change from baseline in St George's Respiratory Questionnaire (domain ‐ symptoms) Show forest plot | 1 | SGRQ units (Fixed, 95% CI) | Totals not selected | |
| 8.1 Partially reversible population (mixed population) | 0 | SGRQ units (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 8.2 Poorly reversible population | 1 | SGRQ units (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 9 Change from baseline in St George's Respiratory Questionnaire (domain ‐ activity) Show forest plot | 1 | SGRQ units (Fixed, 95% CI) | Totals not selected | |
| 9.1 Partially reversible population (mixed population) | 0 | SGRQ units (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 9.2 Poorly reversible population | 1 | SGRQ units (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 10 Change from baseline in St George's Respiratory Questionnaire (domain ‐ impact) Show forest plot | 1 | SGRQ units (Fixed, 95% CI) | Totals not selected | |
| 10.1 Partially reversible population (mixed population) | 0 | SGRQ units (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 10.2 Poorly reversible population | 1 | SGRQ units (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 11 End of treatment St George's Respiratory Questionnaire scores (total score) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 11.1 Partially reversible population (mixed population) | 0 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 11.2 Poorly reversible population | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 12 End of treatment St George's Respiratory Questionnaire scores (domain ‐ breathlessness) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 12.1 Partially reversible population (mixed population) | 0 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 12.2 Poorly reversible population | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 13 Change from baseline in Canadian Respiratory Disease Questionnaire scores Show forest plot | 2 | 712 | Mean Difference (IV, Fixed, 95% CI) | 5.0 [2.48, 7.52] |
| 13.1 Partially reversible population (mixed population) | 2 | 712 | Mean Difference (IV, Fixed, 95% CI) | 5.0 [2.48, 7.52] |
| 13.2 Poorly reversible population | 0 | 0 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 14 Change from baseline in Transitional Dyspnoea Index (TDI) scores Show forest plot | 2 | 707 | Mean Difference (IV, Fixed, 95% CI) | 1.04 [0.56, 1.53] |
| 14.1 Partially reversible population (mixed population) | 2 | 707 | Mean Difference (IV, Fixed, 95% CI) | 1.04 [0.56, 1.53] |
| 14.2 Poorly reversible population | 0 | 0 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 15 End of treatment symptom scores Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 15.1 Partially reversible population (mixed population) | 0 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 15.2 Poorly reversible population | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 16 Change from baseline in predose FEV1 Show forest plot | 5 | Litres (Fixed, 95% CI) | 0.16 [0.14, 0.19] | |
| 16.1 Reversible population | 3 | Litres (Fixed, 95% CI) | 0.19 [0.15, 0.24] | |
| 16.2 Partially reversible population (mixed population) | 0 | Litres (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 16.3 Poorly reversible population | 4 | Litres (Fixed, 95% CI) | 0.15 [0.11, 0.18] | |
| 16.4 Unclear reversibility | 0 | Litres (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 17 Change from baseline in postdose FEV1 Show forest plot | 1 | Litres (Fixed, 95% CI) | Totals not selected | |
| 17.1 Reversible population | 0 | Litres (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 17.2 Partially reversible population (mixed population) | 0 | Litres (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 17.3 Poorly reversible population | 1 | Litres (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 17.4 Unclear reversibility | 0 | Litres (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 18 End of treatment FEV1 (% predicted) Show forest plot | 1 | Std. Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 18.1 Partially reversible population (mixed population) | 0 | Std. Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 18.2 Poorly reversible population | 1 | Std. Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 19 End of treatment FEV1 (Litres) Show forest plot | 2 | Litres (Fixed, 95% CI) | 0.13 [0.10, 0.16] | |
| 19.1 Partially reversible population (mixed population) | 0 | Litres (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 19.2 Poorly reversible population | 2 | Litres (Fixed, 95% CI) | 0.13 [0.10, 0.16] | |
| 20 End of treatment am PEF (L/min) Show forest plot | 1 | L/min (Fixed, 95% CI) | Totals not selected | |
| 20.1 Partially reversible population (mixed population) | 0 | L/min (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 20.2 Poorly reversible population | 1 | L/min (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 21 End of treatment shuttle walk test Show forest plot | 1 | Metres (Fixed, 95% CI) | Totals not selected | |
| 21.1 Partially reversible population (mixed population) | 0 | Metres (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 21.2 Poorly reversible population | 1 | Metres (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 22 End of treatment rescue medication usage (puffs/day) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 22.1 Partially reversible population (mixed population) | 0 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 22.2 Poorly reversible population | 1 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 23 Change from baseline in rescue medication usage (puffs/day) Show forest plot | 2 | 703 | Mean Difference (IV, Fixed, 95% CI) | ‐1.19 [‐1.83, ‐0.55] |
| 23.1 Partially reversible population (mixed population) | 2 | 703 | Mean Difference (IV, Fixed, 95% CI) | ‐1.19 [‐1.83, ‐0.55] |
| 23.2 Poorly reversible population | 0 | 0 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 24 Adverse events ‐ any event Show forest plot | 8 | 5493 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.10 [0.96, 1.27] |
| 24.1 Reversible population | 1 | 126 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.20 [0.59, 2.46] |
| 24.2 Partially reversible population (mixed population) | 2 | 717 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.42 [1.03, 1.96] |
| 24.3 Poorly reversible population | 5 | 4650 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.03 [0.88, 1.21] |
| 25 Adverse events ‐ candidiasis Show forest plot | 6 | 1958 | Odds Ratio (M‐H, Fixed, 95% CI) | 5.73 [3.07, 10.67] |
| 25.1 Reversible population | 1 | 126 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.03 [0.06, 16.88] |
| 25.2 Partially reversible population (mixed population) | 2 | 717 | Odds Ratio (M‐H, Fixed, 95% CI) | 11.13 [3.36, 36.90] |
| 25.3 Poorly reversible population | 3 | 1115 | Odds Ratio (M‐H, Fixed, 95% CI) | 4.40 [2.01, 9.62] |
| 25.4 Unclear reversibility | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 26 Withdrawals due to adverse events Show forest plot | 8 | 5110 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.76 [0.65, 0.88] |
| 26.1 Partially reversible population (mixed population) | 1 | 354 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.69 [0.31, 1.51] |
| 26.2 Poorly reversible population | 6 | 4630 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.76 [0.65, 0.89] |
| 26.3 Unclear reversibility | 1 | 126 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.34 [0.01, 8.47] |
| 27 Withdrawals due to lack of efficacy/exacerbations Show forest plot | 5 | 4620 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.30 [0.21, 0.42] |
| 27.1 Partially reversible population (mixed population) | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 27.2 Poorly reversible population | 5 | 4620 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.30 [0.21, 0.42] |
| 27.3 Unclear reversibility | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 28 Withdrawals Show forest plot | 8 | 5450 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.70 [0.62, 0.78] |
| 28.1 Reversible population | 1 | 121 | Odds Ratio (M‐H, Fixed, 95% CI) | 2.95 [0.30, 29.18] |
| 28.2 Partially reversible population (mixed population) | 2 | 709 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.82 [0.60, 1.13] |
| 28.3 Poorly reversible population | 5 | 4620 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.68 [0.60, 0.76] |
| 28.4 Unclear reversibility | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 29 Adverse events ‐ pneumonia Show forest plot | 7 | 5229 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.80 [1.48, 2.18] |
| 29.1 Reversible population | 1 | 126 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.34 [0.01, 8.47] |
| 29.2 Partially reversible population (mixed population) | 2 | 709 | Odds Ratio (M‐H, Fixed, 95% CI) | 5.55 [0.26, 116.46] |
| 29.3 Poorly reversible population | 4 | 4394 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.80 [1.48, 2.18] |
| 29.4 Unclear reversibility | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 30 Adverse events ‐ nasopharyngitis Show forest plot | 2 | 3535 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.28 [1.05, 1.56] |
| 30.1 Partially reversible population (mixed population) | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 30.2 Poorly reversible population | 2 | 3535 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.28 [1.05, 1.56] |
| 30.3 Unclear reversibility | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 31 Adverse events ‐ pharyngolaryngeal pain Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 31.1 Partially reversible population (mixed population) | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 31.2 Poorly reversible population | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 31.3 Unclear reversibility | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 32 Adverse events ‐ upper respiratory tract infection Show forest plot | 5 | 4963 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.23 [1.04, 1.47] |
| 32.1 Partially reversible population (mixed population) | 2 | 709 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.25 [0.81, 1.92] |
| 32.2 Poorly reversible population | 3 | 4254 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.23 [1.02, 1.48] |
| 32.3 Unclear reversibility | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 33 Adverse events ‐ headache Show forest plot | 5 | 4367 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.06 [0.85, 1.32] |
| 33.1 Reversible population | 1 | 123 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.22 [0.02, 2.01] |
| 33.2 Partially reversible population (mixed population) | 2 | 709 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.38 [0.91, 2.10] |
| 33.3 Poorly reversible population | 2 | 3535 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.97 [0.75, 1.27] |
| 33.4 Unclear reversibility | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 34 Adverse events ‐ hoarseness Show forest plot | 2 | 585 | Odds Ratio (M‐H, Fixed, 95% CI) | 8.79 [1.11, 69.62] |
| 34.1 Partially reversible population (mixed population) | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 34.2 Poorly reversible population | 2 | 585 | Odds Ratio (M‐H, Fixed, 95% CI) | 8.79 [1.11, 69.62] |
| 34.3 Unclear reversibility | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 35 Adverse events ‐ pyrexia Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 35.1 Partially reversible population (mixed population) | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 35.2 Poorly reversible population | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 35.3 Unclear reversibility | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 36 Adverse events ‐ cough Show forest plot | 4 | 1057 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.61 [0.29, 1.29] |
| 36.1 Reversible population | 1 | 126 | Odds Ratio (M‐H, Fixed, 95% CI) | 3.15 [0.13, 78.72] |
| 36.2 Partially reversible population (mixed population) | 1 | 346 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.49 [0.18, 1.31] |
| 36.3 Poorly reversible population | 2 | 585 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.66 [0.18, 2.44] |
| 36.4 Unclear reversibility | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 37 Adverse events ‐ palpitations Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 37.1 Partially reversible population (mixed population) | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 37.2 Poorly reversible population | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 37.3 Unclear reversibility | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 38 Adverse events ‐ mouth ulceration Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 38.1 Partially reversible population (mixed population) | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 38.2 Poorly reversible population | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 38.3 Unclear reversibility | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 39 Adverse events ‐ toothache Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 39.1 Partially reversible population (mixed population) | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 39.2 Poorly reversible population | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 39.3 Unclear reversibility | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 40 Adverse events ‐ urinary tract infection Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 40.1 Partially reversible population (mixed population) | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 40.2 Poorly reversible population | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 40.3 Unclear reversibility | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 41 Adverse events ‐ dyspnoea Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 41.1 Partially reversible population (mixed population) | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 41.2 Poorly reversible population | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 41.3 Unclear reversibility | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 42 Adverse events ‐ blood glucose increased Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 42.1 Partially reversible population (mixed population) | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 42.2 Poorly reversible population | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 42.3 Unclear reversibility | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 43 Adverse events ‐ insomnia Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 43.1 Partially reversible population (mixed population) | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 43.2 Poorly reversible population | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 43.3 Unclear reversibility | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 44 Adverse events ‐ bronchitis Show forest plot | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 44.1 Partially reversible population (mixed population) | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 44.2 Poorly reversible population | 1 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 44.3 Unclear reversibility | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Severe Exacerbations Show forest plot | 2 | Rate ratio (Fixed, 95% CI) | 0.74 [0.62, 0.88] | |
| 1.1 Partially reversible | 0 | Rate ratio (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 1.2 Poorly reversible | 2 | Rate ratio (Fixed, 95% CI) | 0.74 [0.62, 0.88] | |
| 2 Mean severe exacerbation rates per patient per year Show forest plot | 2 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 2.1 Partially reversible population (mixed population) | 0 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 2.2 Poorly reversible population | 2 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3 Quality of life ‐ SGRQ (change scores) Show forest plot | 2 | Mean Difference (IV, Fixed, 95% CI) | Totals not selected | |
| 3.1 Partially reversible population (mixed population) | 0 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 3.2 Poorly reversible population | 2 | Mean Difference (IV, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4 Quality of life ‐ change scores Show forest plot | 2 | SGRQ (Fixed, 95% CI) | ‐6.06 [‐7.90, ‐4.22] | |
| 4.1 Partially reversible (mixed population) | 0 | SGRQ (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 4.2 Poorly reversible | 2 | SGRQ (Fixed, 95% CI) | ‐6.06 [‐7.90, ‐4.22] | |
| 5 Rescue medication usage Show forest plot | 2 | Puffs per day (Fixed, 95% CI) | ‐0.87 [‐1.15, ‐0.58] | |
| 5.1 Partially reversible (mixed population) | 0 | Puffs per day (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 5.2 Poorly reversible | 2 | Puffs per day (Fixed, 95% CI) | ‐0.87 [‐1.15, ‐0.58] | |
| 6 Symptoms (change scores) Show forest plot | 2 | Symptom scale (Fixed, 95% CI) | ‐0.63 [‐0.90, ‐0.37] | |
| 6.1 Partially reversible (mixed population) | 0 | Symptom scale (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 6.2 Poorly reversible | 2 | Symptom scale (Fixed, 95% CI) | ‐0.63 [‐0.90, ‐0.37] | |
| 7 Mean FEV1 (% change from baseline) Show forest plot | 2 | % increase (Fixed, 95% CI) | 14.40 [11.91, 16.90] | |
| 7.1 Partially reversible (mixed population) | 0 | % increase (Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 7.2 Poorly reversible | 2 | % increase (Fixed, 95% CI) | 14.40 [11.91, 16.90] | |
| 8 Adverse events ‐ 'serious events' Show forest plot | 2 | 923 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.06 [0.78, 1.45] |
| 8.1 Partially reversible population (mixed population) | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 8.2 Poorly reversible population | 2 | 923 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.06 [0.78, 1.45] |
| 9 Adverse events ‐ candidiasis | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | Totals not selected | |
| 9.1 Partially reversible population (mixed population) | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 9.2 Poorly reversible population | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] | |
| 10 Adverse events ‐ pneumonia Show forest plot | 1 | 510 | Odds Ratio (M‐H, Fixed, 95% CI) | 4.13 [0.87, 19.64] |
| 11 Withdrawals due to worsening COPD symptoms Show forest plot | 2 | 923 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.40 [0.28, 0.58] |
| 11.1 Partially reversible population (mixed population) | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 11.2 Poorly reversible population | 2 | 923 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.40 [0.28, 0.58] |
| 12 Withdrawals due to adverse events Show forest plot | 2 | 923 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.30 [0.78, 2.18] |
| 12.1 Partially reversible population (mixed population) | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 12.2 Poorly reversible population | 2 | 923 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.30 [0.78, 2.18] |
| 13 Mortality Show forest plot | 2 | 923 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.78 [0.35, 1.73] |
| 13.1 Mortality as primary outcome | 0 | 0 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 13.2 Mortality data collected as secondary/unpublished outcome | 2 | 923 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.78 [0.35, 1.73] |
