Background

The mortality, morbidity and economic burden of chronic obstructive pulmonary disease (COPD) is well documented [1, 2]. Dyspnea as a cardinal symptom of COPD is a major cause of disability and anxiety associated with the disease, prompting the medical community to pursue effective treatments for the relief of dyspnea [3]. Recent guidelines recommend the regular use of inhaled long-acting bronchodilators to alleviate dyspnea in patients with symptomatic disease, with the addition of inhaled corticosteroids for patients who experience repeated exacerbations [1, 2].

Indacaterol as a novel, once-daily (od), inhaled, long-acting ß2-agonist provides sustained bronchodilation for patients with moderate to severe COPD, with a rapid onset following the first dose [4, 5]. It has been recently approved in Europe at doses of 150 μg and 300 μg od, in the United States at 75 μg od, and in China at 150 μg od for maintenance treatment of airflow limitation in patients with COPD. A number of randomized clinical trials suggest that indacaterol may improve dyspnea in patients with stable COPD, as indicated by changes in transition dyspnea index (TDI) [510]. However, all of these trials were statistically powered on forced expiratory volume in 1 second (FEV1) as a primary endpoint. The outcome measure of TDI was included as one of the secondary endpoints in these trials, and so might not have been adequately powered.

Therefore, we undertook a systematic review and meta-analysis of available randomized placebo-controlled studies to assess the efficacy of indacaterol on the important clinical outcome of dyspnea. The objectives of this meta-analysis were to combine data from existing randomized placebo-controlled trials, to use the number of patients achieving the minimum clinically important difference (MCID) for TDI score ≥ 1 as an outcome measure, and to evaluate the efficacy of once-daily indacaterol of licensed 75 μg, 150 μg, or 300 μg doses relative to placebo on dyspnea in patients with stable COPD.

Methods

Data sources and selection criteria

We identified published studies between January 2007 and May 2012 from MEDLINE, EMBASE, and the Cochrane Controlled Trials Register (CENTRAL) databases using the terms indacaterol, long-acting ß 2 -agonist AND chronic obstructive pulmonary disease. We performed a search of relevant files from the Novartis trials results database (http://www.novctrd.com/ctrdWebApp/clinicaltrialrepository/public/login.jsp). We also performed a manual search of references cited in published original and review articles, and in clinical practice guidelines. Trials published solely in abstract form were excluded because they contain preliminary and rudimentary information and may not provide enough details to allow full analysis. Two reviewers (JNH and LD) then independently screened potentially relevant trials from titles and abstracts. Using the full texts as necessary, the two reviewers independently identified eligible articles for full review. Finally, we reviewed eligible articles to determine whether they qualified for meta-analysis. Differences were resolved by discussion.

To be included, studies had to meet all the following criteria: a) target population of stable COPD consistent with American Thoracic Society/European Respiratory Society [11] or Global Initiative for Chronic Obstructive Lung Disease (GOLD) diagnostic criteria [12]; b) randomized placebo-controlled trials comparing indacaterol 75 μg, 150 μg, or 300 μg od with placebo; c) studies that followed patients for 12 weeks or more after randomization; d) studies that included assessment of dyspnea by transition dyspnea index (TDI) in outcome measures.

Data extraction

Two reviewers independently read each article that met inclusion criteria and performed data extraction using a pre-designed data collection form. Missing data were obtained from the manufacturer. Disagreement and uncertainty were solved by discussion. Consensus was reached for all data. Data extracted from each article included: first author’s name and year of publication; study design; treatment arms; number of patients; treatment duration; baseline clinical characteristics including age, gender, smoking history, post-bronchodilator spirometry, and total score of the baseline dyspnea index (BDI). The outcome measure of TDI was recorded as total score and as the number of patients with TDI ≥ 1.

Baseline dyspnea index (BDI) and transition dyspnea index (TDI)

Dyspnea is often measured using the BDI and TDI, a tool recommended by regulatory authorities for inclusion in clinical trials of treatments for COPD [13]. The BDI and TDI, as multidimensional instruments, each has three domains: functional impairment, magnitude of task, and magnitude of effort [14]. The BDI domains measure baseline dyspnea severity, and are rated from 0 (severe) to 4 (unimpaired) and summed to provide a BDI total score of 0 to 12, with a lower score indicating more severe dyspnea. The TDI domains measure change from the baseline dyspnea index (BDI) over time, rated on a scale of +3 (major improvement) to −3 (major deterioration). The TDI has been shown to be valid, reliable and responsive [14, 15]. The minimum clinically important difference (MCID) for the TDI is an improvement from the BDI score of ≥1 unit [16, 17].

Statistical analyses

Odds ratio (OR) for likelihood of achieving TDI score ≥1 after 12 weeks of treatment was used as a measure to compare indacaterol relative to placebo. We calculated pooled ORs with the DerSimonian-Laird random effects model [18], usually regarded as more appropriate than other statistical approaches when potential heterogeneity is present between studies [19, 20]. We performed separate analyses for indacaterol 75 μg versus placebo, indacaterol 150 μg versus placebo, and indacaterol 300 μg versus placebo. We calculated the 95% confidence intervals around the ORs, and assessed heterogeneity across studies with the chi-square test and I 2 (p < 0.10, I 2 > 25%). We then created forest plots of the individual studies and combined estimates. All analyses were performed with meta-analysis software (MetaAnalyst version beta 3.13, Tufts Medical Center, Boston, Massachusetts).

Results

Characteristics of included studies

Six studies met the inclusion criteria [59], the designs of which are summarized in Table 1. The studies varied from 12 to 52 weeks in duration, and so to permit comparison across trials, assessments at 12 weeks of treatment were used for the 52-week study [6] and for two 26-week studies [5, 7]. We did not consider data from tiotropium, formoterol, and salmeterol arms for the current meta-analysis, since data for each of these arms would be provided from one study.

Table 1 Summary of study designs for the studies included in this meta-analysis

Table 2 illustrates patient populations and baseline characteristics. Of the six studies providing the data for the meta-analysis, four had similar inclusion criteria [57, 9], recruiting male and female patients aged ≥40 years with a clinical diagnosis of moderate-to-severe COPD as per the GOLD 2005 criteria and a smoking history of ≥20 pack-years. Postbronchodilator FEV1 was to be <80% and ≥30% predicted and post-bronchodilator FEV1/forced vital capacity <70% [57, 9]. The two identical indacaterol 75 μg trials enrolled patients with moderate-to-severe COPD defined at that time using GOLD 2008 criteria, aged ≥40 years and with a smoking history of ≥10 pack-years [8].

Table 2 Patient characteristics at baseline in the individual studies

Dyspnea – BDI and TDI

In all trials included in the meta-analysis, dyspnea was measured at baseline using the BDI. As illustrated in Table 2, patients had moderate severity of dyspnea at baseline with mean BDI total scores ranging from 5.81 to 7.67. After 12 weeks of treatment, dyspnea was measured using the TDI (Table 3), which captured changes from baseline. Data are presented as mean TDI total scores and as the number of patients with TDI score ≥1 unit. Results were analyzed for the number of patients responding with a change of TDI equal to or greater than the MCID (‘responder analysis’) (Table 3).

Table 3 Indacaterol on dyspnea as measured by the TDI

Indacaterol 75 μg versus placebo

Two randomized placebo-controlled studies had identical entry criteria and study designs, which compared indacaterol 75 μg once daily with placebo after 12 weeks of treatment [8]. A meta-analysis that combined the two studies produced a pooled OR estimate of 1.784 (95% CI 1.282–2.482) with no evidence of heterogeneity (P = 0.474, I2 = 0.000), indicating that relative to placebo, patients receiving indacaterol 75 μg are more likely to achieve TDI score ≥1 after 12 weeks of treatment. Figure 1 shows a forest plot of OR estimates from these studies.

Figure 1
figure 1

Forest plot of odds ratios and 95% confidence intervals. Relative to placebo, patients receiving indacaterol 75 μg once daily are more likely to achieve TDI score equal to or greater than one unit.

Indacaterol 150 μg versus placebo

Three randomized placebo-controlled trials compared indacaterol 150 μg once daily with placebo [5, 7, 9]. A meta-analysis that combined the three studies produced a pooled OR estimate of 2.149 (95% CI 1.746–2.645) with no evidence of heterogeneity (P = 0.686, I2 = 0.000), favoring patients who received indacaterol 150μg once daily. Figure 2 shows a forest plot of OR estimates from these studies.

Figure 2
figure 2

Meta-analysis of three randomized trials compared indacaterol 150 μg once daily with placebo. Relative to placebo, patients receiving indacaterol 150 μg once daily are more likely to achieve TDI score equal to or greater than one unit.

Indacaterol 300 μg versus placebo

Three trials compared indacaterol 300 μg once daily with placebo [5, 6, 9]. Figure 3 shows a forest plot of OR estimates from these studies. The combined OR estimate was 2.458 (95% CI 2.010–3.006) with no evidence of heterogeneity (P = 0.525, I2 = 0.000), again favoring patients who received indacaterol 300 μg once daily.

Figure 3
figure 3

Meta-analysis of three randomized trials compared indacaterol 300 μg once daily with placebo. Relative to placebo, patients receiving indacaterol 300 μg once daily are more likely to achieve TDI score equal to or greater than one unit.

Discussion

The present meta-analyses combined data from existing randomized placebo-controlled trials, using number of patients achieving the minimum clinically important difference (MCID) for TDI score ≥1 as an outcome measure, and evaluated the efficacy of once-daily indacaterol on TDI scores in patients with stable COPD. A favorable effect was consistently obtained for indacaterol over placebo: 75 μg od OR 1.784 (95% CI 1.282 to 2.482); 150 μg od OR 2.149 (95% CI 1.746 to 2.645); and 300 μg od OR 2.458 (95% CI 2.010 to 3.006). A trend of increasing patient benefit was observed as indacaterol doses increased.

Recent guidelines recommend regular use of bronchodilators, such as indacaterol, as a long-term maintenance treatment of airflow limitation in COPD [1]. The extent to which indacaterol improves airflow in patients with COPD is well studied [4, 5, 2124]. In these phase III studies, indacaterol provided 24-h bronchodilation on once-daily dosing with an effect that was sustained during treatment for up to one year. However, from the patients’ viewpoint, it may be more important to know whether this therapy improves dyspnea associated with daily activities, and these meta-analyses of available randomized placebo-controlled trials comparing indacaterol with placebo allows us to examine the efficacy of indacaterol on this important patient outcome. In all trials included in the meta-analyses, dyspnea was assessed using validated multidimensional instruments of BDI and TDI that are widely used to measure treatment effects in COPD [1317]. When data were analyzed for a pre-determined outcome, i.e. the number of patients responding with a change of TDI equal to or greater than the minimum clinically important difference MCID (a ‘responder analysis’), a favorable effect was obtained for all three doses of indacaterol. The overall OR estimates of the meta-analyses are largely consistent with the results from the individual studies, in which the ORs for patients achieving a MCID in the indacaterol group compared with the placebo group varied from 1.582 to 2.015 for the 75 μg dose (Figure 1) [8], 1.950 to 2.303 for the 150 μg dose (Figure 2) [5, 7], and 2.520 to 2.626 for the 300 μg dose (Figure 3) [5, 6], with a exception of the Asian study [9]. It becomes apparent that patients receiving indacaterol had clinically significant improvements in symptoms of dyspnea compared to placebo over 12 weeks of treatment.

The overall odds ratio for response in TDI appeared to vary depending on indacaterol doses, and tended to increase with increasing indacaterol doses. Additionally, the percentage of patients who exceeded the one unit of the MCID in TDI varied in different indacaterol doses. The percentage in the analysis of indacaterol 75 μg once daily was 48%, compared with 34% in placebo. Compared to placebo, a higher percentage of patients achieved the MCID with the indacaterol 150 μg (60% versus 41%) and indacaterol 300 μg doses (63% versus 41%). Renard and his colleagues recently performed a model based analysis of the bronchodilatory dose response to indacaterol in patients with COPD [25]. The analysis demonstrated that indacaterol dosages of 75 μg once daily and above achieved minimal clinically important improvements in predicted trough FEV1 response, although dosages of 150 μg and 300 μg once daily provided optimum bronchodilation. The analysis also demonstrated that disease severity, as determined by FEV1, significantly affected dose response, suggesting that higher doses may be required in patients with more severe COPD to achieve optimal reduction of dyspnea.

In six trials included in the meta-analysis (n=5405), patients with COPD who received indacaterol 75 μg, or 150 μg, or 300 μg od had a significantly higher trough FEV1 than placebo after at least 12 weeks, with indacaterol increasing trough FEV1 by 120ml to 200 ml over placebo at week 12 [59]. The overall improvement in FEV1, though modest, may be sufficient to decrease the extent of hyperinflation which contributes to the sensation of dyspnea [2629]. A better indicator of the effect of a bronchodilator on hyperinflation (and therefore the sensation of dyspnea) is perhaps inspiratory capacity. One could hypothesize that the prolonged bronchodilation observed with indacaterol would be associated with reductions of air trapping, and therefore reductions in hyperinflation, which would then be reflected in improvements in the sensation of dyspnea. There are relatively few indacaterol studies that included both an assessment of inspiratory capacity and of dyspnea, however, two short-term exercise studies comparing the 300 μg dose of indacaterol with placebo have been published [30, 31]. The BDI/TDI was included in one of the studies, an improvement of 182 ml in resting inspiratory capacity after 14 days (p<0.05 vs. placebo) was associated with a change of TDI total score of +3.33 (p<0.01 vs. placebo) [31], providing indirect evidence to support the hypothesis.

Conclusions

Once daily indacaterol provides clinically significant improvements in dyspnea compared with placebo after 12 weeks of treatment. Incremental benefits in TDI were observed with increasing doses of indacaterol. Indacaterol may provide patients and physicians with a useful treatment option in symptomatic patients with dyspnea. Further studies, that investigate the relationship between disease severity as determined by FEV1 and dose response to indacaterol in terms of dyspnea, are needed to specifically address the question whether higher doses are required in patients with more severe COPD to achieve optimal reduction of dyspnea.