The goals of asthma management from a healthcare professional’s perspective are to control symptoms, restore or maintain normal airway function, prevent exacerbations and maintain normal activity levels.[1] Clinical measures of respiratory function, such as peak expiratory flow (PEF), evaluation of symptoms and exacerbations are important outcomes used in the challenge of meeting these goals. From the patient’s perspective, however, the primary aim of asthma therapy is essentially to improve health-related quality of life (HR-QOL). Although symptoms and lung function measurements may be of concern to patients, these measurements do not indicate the extent to which asthma may affect everyday lives of these patients, because there is only a weak association between clinical and HR-QOL outcomes.[2–6]

Measurement of HR-QOL may reveal beneficial treatment effects not captured by objective monitoring, and so can complement clinical and physiological assessments.[7] Therefore, improving HR-QOL and thus allowing patients to participate more fully in everyday life should be a major aim of asthma management.

There is growing evidence to suggest that for patients whose asthma is uncontrolled on inhaled corticosteroids, it is clinically more effective[8–11] and more cost-effective[12,13] to add a long-acting β2-adrenoceptor agonist (β2-agonist) to their therapy than to double the corticosteroid dose. Indeed, this treatment step is now recommended in the new Global Initiative for Asthma (GINA) guidelines.[1] A combination of long-acting β2-agonist and inhaled corticosteroid has been introduced (salmeterol/ fluticasone propionate). Studies have shown that salmeterol/ fluticasone propionate administered via a dry powder inhaler (Diskus™/Accuhaler™) is more effective than fluticasone alone[14] and is clinically equivalent to fluticasone propionate and salmeterol co-administered in separate inhalers.[15]

In this investigation, we present one of the first comparisons of the impact on HR-QOL of a long-acting β2-agonist/inhaled corticosteroid combination versus a higher dose of inhaled corticosteroid alone. This study was conducted as part of a double-blind (double-dummy) randomized controlled trial designed to compare the safety and efficacy of the combined product, salmeterol 50μg plus fluticasone propionate 250μg with an increased dose of inhaled corticosteroid alone, budesonide 800μg, both given twice daily to patients with moderate-to-severe asthma previously uncontrolled on inhaled corticosteroids. This study is also the first HR-QOL study to compare a long-acting β2-agonist/ inhaled corticosteroid combination with another inhaled corticosteroid alone, administered via different dry powder inhalers. The efficacy and safety results have been published elsewhere.[16]

Methods

Subjects

Male and female patients, aged ≥12 years, with a documented history of reversible airways obstruction were recruited. Eligible patients had received inhaled beclomethasone dipropionate or budesonide at doses of 800–1200 μg/day or fluticasone propionate 400–600 μg/day for at least 4 weeks. Patients were excluded if in the 4 weeks before the run-in period they had experienced a lower respiratory tract infection or acute exacerbation of reversible airways obstruction (requiring hospitalization), or had changed their asthma medication or received oral, parenteral or depot corticosteroids. Other exclusion criteria included a history of smoking ≥10 pack years and treatment with a long-acting β2-agonist or slow-release bronchodilator in the 2 weeks before the run-in period. All patients gave written informed consent. The study was conducted in accordance with the Declaration of Helsinki, and local ethics committee approval was obtained at each center.

Study Design

This was a multicenter, double-blind (double-dummy), randomized, parallel-group trial comparing the efficacy and tolerability of inhaled salmeterol/fluticasone propionate and budesonide in patients with moderate to severe persistent asthma uncontrolled by inhaled corticosteroids. Although the total treatment duration was 24 weeks, HR-QOL was assessed at baseline and after 12 weeks of treatment (or earlier upon study withdrawal) at 26 centers in Belgium, the Netherlands, Australia and South Africa.

Eligible patients underwent a 2-week run-in period, during which they used their usual inhaled corticosteroid medication. Inhaled albuterol (salbutamol) was provided for use as required but all other anti-asthma therapies were discontinued. At the end of this period, patients were required to fulfill the criteria summarized in table I before commencing the treatment period. Those patients who failed to complete daily record cards correctly or measure and record PEF values were excluded. The remainder were randomized to receive twice daily treatment with either salmeterol/fluticasone propionate 50/250μg (Seretide™/ Advair™) via a Diskus™inhaler plus placebo via a Turbuhaler™ (total daily dose: salmeterol 100μg plus fluticasone propionate 500μg), or budesonide 800μg (Pulmicort™) via a Turbuhaler™ plus placebo via a Diskus™(total daily dose: budesonide 1600μg). Patients completed the self-administered version of the Asthma Quality of Life Questionnaire (AQLQ) at baseline and after 12 weeks of treatment (or at study withdrawal).

Table I
figure Tab1

Patient inclusion criteriaa

Full size image

Quality-Of-Life Assessment

The AQLQ is a well-validated, disease-specific instrument designed to measure HR-QOL in clinical trials. Previous studies have shown the AQLQ to have good reliability, responsiveness and longitudinal construct validity.[2,5,17,18] The AQLQ comprises 32 questions (items) categorized into four domains: activity limitation (11 items); asthma symptoms (12 items); emotional functioning (5 items); and exposure to environmental stimuli (4 items). In order to individualize the assessment, the first five items within the activity limitation domain are self-selected by patients at baseline to reflect their own circumstances and values. These items are then retained for subsequent administrations of the questionnaire. Each item is scored using a 7-point scale, from 1 (severe impairment) to 7 (no impairment). In this study, the self-administered version of the AQLQ was used.

Each domain of the AQLQ is scored by calculating the mean value of the responses to its constituent items. The overall AQLQ score for each patient is calculated as the mean value of all 32 items.

Data Analysis

Data from patients who completed questionnaires at both baseline and endpoint (either at week 12 or upon early withdrawal) were analyzed. Results were summarized by treatment group as mean scores ± standard deviations.

Changes from baseline within each treatment group were analyzed using paired t-tests. Changes that were statistically significant were evaluated for clinical importance; a change in the mean overall or domain score of 0.5 has been shown to represent the smallest change of importance to the patient.[19]

An analysis of covariance (ANCOVA) model was used to compare the changes in AQLQ scores (measured as least square mean ± standard error of the mean [SEM]) between the two treatment groups. The country and baseline scores were covariates in this model. Treatment and cluster effects were evaluated in a preliminary model.

Although assessment of between-group differences based on mean score changes of ≥0.5 is important, it does not provide complete information on the relative benefits of treatments.[20] Because of the heterogeneity of responses from patients, it is important to assess not only the group mean, but also the distribution of responses about this mean. Accordingly, in addition to evaluating mean differences using an ANCOVA model, the changes in AQLQ scores were analyzed using a number-needed-to-treat (NNT) analysis that has previously been applied to AQLQ data.[20,21]

A threshold score change of 0.5 (representing a clinically important change) was used to determine the proportion of patients in each treatment group who experienced improvement (decrease in score of ≥0.5 from baseline), deterioration (score increase of ≥0.5) or no change (score change of −0.49 to +0.49) in their overall AQLQ scores. Assuming the responses to both treatments to be independent, these patient proportions were subjected to NNT analysis which provided an estimate of the number of patients who would need to be treated with a given intervention for one patient to have a clinically meaningful improvement in HR-QOL, relative to improvements seen with the comparator. NNT is therefore a means of assessing the incremental gains in HR-QOL that may be achieved with a more beneficial treatment.

Results

In total, 144 patients completed the AQLQ at baseline. Of these, 113 (61 male, 52 female) also completed a questionnaire either at the end of treatment (n=110) or upon early withdrawal from the study (salmeterol/fluticasone propionate=1, budesonide=2). One patient treated with budesonide experienced an asthma exacerbation, the other two patients withdrew for non-medical reasons. Thirty-one patients failed to complete the AQLQ either at the end of the study (n=10) or at premature withdrawal (n=21) and therefore could not be included in the analysis. Of the 21 patients who withdrew, 12 patients were in the salmeterol/fluticasone group (one patient withdrew because of asthma exacerbation, two had adverse events not related to study medications and, nine withdrew for non-medical reasons). Nine of the 21 patients who withdrew were in the budesonide group (three patients withdrew because of exacerbation/lack of efficacy and six for non-medical reasons). Of the113 patients who completed an endpoint AQLQ, 55 were randomized to receive salmeterol/fluticasone propionate and 58 to receive budesonide. The baseline and demographic characteristics of the two groups were similar (table II). The mean age of the patients was 50 years (range 16–80 years).

Table II
figure Tab2

Demographic characteristics of patients at baselinea

Full size image

Within-Group Comparisons

Mean overall AQLQ scores increased significantly from baseline to treatment endpoint (either at the 12-week visit or early withdrawal) in both the salmeterol/fluticasone propionate and the budesonide groups (table III). In both treatment groups, mean changes from baseline were statistically significant in each of the four AQLQ domains. The overall score increase in both treatment groups exceeded the 0.5 threshold for a clinically important improvement in asthma quality of life and in all four domains for patients treated with salmeterol/fluticasone propionate but in only two domains for patients treated with budesonide.

Table III
figure Tab3

Least square mean changes from baseline in Asthma Quality of Life Questionnaire (AQLQ) scores following treatment

Full size image

Between-Groups Comparison

The differences between the AQLQ score changes in the two treatment groups are shown in table III. The mean improvement in overall AQLQ score in the salmeterol/fluticasone propionate group was significantly greater than in the budesonide group (mean difference=0.45; SEM±0.14, p=0.002). Improvements in all AQLQ domain scores were significantly greater in the salmeterol/fluticasone propionate group than the budesonide group (p <0.05, ANCOVA) [figure 1]. The largest difference between the improvements in the two groups, both in favor of salmeterol/fluticasone propionate, occurred in the symptoms and emotional function domains (mean and SEM of 0.54±0.17 and 0.51±0.17, respectively; p <0.005).

Fig. 1
figure 1

Between-group differences in least square mean change (±95% confidence limits) from baseline in overall and domain Asthma Quality of Life Questionnaire (AQLQ) scores at treatment endpoint (12 weeks or early withdrawal) in patients with asthma who received inhaled salmeterol/fluticasone propionate 50/250μg twice daily or budesonide 800μg twice daily. Differences between treatments [salmeterol/fluticasone propionate [SFC] minus budesonide [BUD]) were analyzed using analysis of covariance. * p <0.05; ** p <0.01.

Full size image

Number-Needed-To-Treat Analysis

Approximately 70% of patients in the salmeterol/fluticasone propionate group experienced an improvement in HR-QOL (a decrease in overall AQLQ score of ≥0.5), 30% remained unchanged (change in score of between −0.49 and 0.49) and none of the patients deteriorated (an increase in score of ≥0.5). In the budesonide group, 43% of patients improved, 45% remained unchanged and 12% deteriorated. These patient proportions resulted in an NNT of 3.4 for the overall score. For the individual domains, NNT values ranged from 4.0 for exposure to environmental stimuli to 7.3 for emotional functioning (table IV).

Table IV
figure Tab4

Matrix showing the proportions of patients in each treatment group who experienced clinically important changes in overall or domain Asthma Quality of Life Questionnaire (AQLQ) scores compared with baseline, and the number-needed-to-treat (NNT) value

Full size image

Discussion

HR-QOL is an outcome measure that complements clinical and physiological assessments. It is an important measure as it evaluates the impact of asthma treatments from the patients’ perspective. Although salmeterol and fluticasone propionate have both been shown to improve HR-QOL in patients with asthma,[6,22] the present study is one of the first to evaluate the impact of the salmeterol/fluticasone propionate combination product on HR-QOL. Furthermore, this study is one of the first to compare the impact of alternative treatment options on the HR-QOL of patients. Asthma guidelines recommend that, for patients whose asthma is uncontrolled on current doses of inhaled corticosteroids, two potential treatment options are to increase the corticosteroid dose or add-on therapy with a long-acting bronchodilator. The results demonstrate that for patients with moderate-to-severe asthma uncontrolled on inhaled corticosteroids (beclomethasone dipropionate 800–1200μg or equivalent), treatment with salmeterol/fluticasone propionate 50/250μg twice daily resulted in significantly greater improvements in their HR-QOL than treatment with an increased dose of inhaled corticosteroid alone (budesonide 800μg twice daily). These effects on HR-QOL are consistent with the superior clinical efficacy of salmeterol/fluticasone propionate demonstrated in the clinical component of this study.[16]

The AQLQ was designed to assess aspects of asthma considered important by patients themselves. To aid interpretation of AQLQ data, overall and domain score changes have been related to patients’ perceptions of HR-QOL improvement and deterioration.[19] By this method, the smallest AQLQ change of importance to the patient (overall and in each domain) has been identified as 0.5. Between-group comparisons revealed that improvements in quality of life in the salmeterol/fluticasone group were significantly greater than in the budesonide group for overall AQLQ score and all individual domains. Although the between-group difference was greater than 0.5 for the symptom (0.54) and emotional function (0.51) domains, it did not reach this value for the overall score (0.45) and the activity (0.33) and environmental (0.41) domains. In the past, it might have been interpreted that the symptom and emotional function domains showed clinically important improvements whereas the overall score and other domains did not. However, recent work on the interpretation of continuous data has shown that basing interpretation only on mean values may be very erroneous and that interpretation should also take into account the heterogeneity of patients’ responses to interventions; in other words, the variance about the mean.[21] Having shown that a difference between two treatments is unlikely to have occurred by chance alone (p<0.05) expressing the results in terms of the NNT, calculated from the proportion of patients improving and deteriorating by more than the minimal important difference (0.5) in each group, not only takes into account the variance but presents the results in a way that is easy to understand. It provides the number of patients who need to be treated with the preferred treatment for one patient to benefit compared with the alternative treatment. The analysis of this study showed that the NNT for the overall score of the AQLQ was 3.4. This means that between three and four patients need to be treated with salmeterol/fluticasone propionate for one patient to have a clinically important improvement in asthma-specific quality of life compared with treatment with budesonide. The NNTs for each of the four domains were similar (activities 6.1, symptoms 5.0, emotions 7.3 and, environment 4.0). In order to place these NNT results in context, it is useful to compare the values obtained in this study with those calculated in other published studies that used the AQLQ. In a comparison of salmeterol and albuterol, the NNT in favor of salmeterol was 4.5.[21] A comparison of beclomethasone dipropionate delivered by conventional versus extra fine aerosol produced an NNT value of 7.3 in favor of the extra fine aerosol.[7] In the Formoterol And Corticosteroids Establishing Therapy (FACET) study, the addition of the long-acting β2-agonist formoterol to budesonide compared with an increased dose of budesonide, resulted in an NNT of 11.9 in favor of the β2-agonist.[20] Finally, a study comparing the effects of equipotent dosages of high-dose inhaled corticosteroids in severe asthma reported an NNT of 3.6 for overall AQLQ score for fluticasone propionate relative to budesonide or beclomethasone dipropionate.[23] In the context of these four studies, the low NNT values observed in the present study suggest that salmeterol/ fluticasone propionate is an effective intervention for improving HR-QOL in patients with asthma when compared with increased doses of budesonide.

In summary, twice-daily treatment with salmeterol/fluticasone propionate (salmeterol/fluticasone propionate 50/250μg combination product) resulted in clinically important improvements in patients’ asthma-specific quality of life that were greater than those produced by twice daily treatment with budesonide 800μg. This study supports previous evidence that in patients with asthma, uncontrolled on their current dosages of inhaled corticosteroids, it is more beneficial to treat patients with an additional long-acting β2-agonist, such as salmeterol, than to use a higher dose of inhaled corticosteroid.