Fluticasone furoate (FF)/vilanterol (100/25 mcg or 200/25 mcg) or FF (100 mcg) in persistent asthma.

OBJECTIVES
Fluticasone furoate (FF; inhaled corticosteroid) combined with vilanterol (VI; long-acting beta(2) agonist) is a once-daily therapy for asthma and chronic obstructive pulmonary disease. This 12-week phase III study compared the efficacy and safety of once-daily (evening dosing) FF/VI 100/25 mcg versus FF 100 mcg (primary objective) and FF/VI 100/25 mcg versus FF/VI 200/25 mcg (descriptive comparison only) in patients (n = 1039) ≥12 years with moderate-to-severe persistent asthma.


METHODS
The primary end point was weighted mean (wm) 0-24-h serial forced expiratory volume in 1 s (FEV(1)) at week 12. Secondary end points (change from baseline) were trough FEV(1) and the proportion (%) of rescue-free 24-h periods (both powered), the proportion (%) of symptom-free 24-h periods, and morning and evening peak expiratory flow (PEF). Safety data (adverse events, AEs) were collected throughout.


RESULTS
Compared with FF 100 mcg, FF/VI 100/25 mcg significantly improved wmFEV(1) (p < 0.001), trough FEV(1) (p = 0.014), % rescue-free (p < 0.001), % symptom-free (p = 0.002) 24-h periods, and morning and evening PEF (p < 0.001). FF/VI 200/25 mcg produced small numerical improvements versus FF/VI 100/25 mcg for all end points. Incidence of AEs was similar across groups.


CONCLUSIONS
FF/VI 100/25 mcg resulted in significant improvements in all primary and secondary end points versus FF 100 mcg. Numerical improvements occurred with FF/VI 200/25 mcg versus FF/VI 100/25 mcg. All treatments were well tolerated.


Introduction
Inhaled corticosteroids (ICS) are the mainstay of the treatment for asthma [1][2][3]; their introduction in the 1970s represented a leap forward in the management of asthma. However, the role of inflammation in asthma was not fully understood until the 1990s, when guidelines were developed that recommended ICS as first-line treatment for patients with asthma not controlled with intermittent short-acting beta 2 agonist (SABA) use. For patients whose asthma symptoms persist when treated with ICS, addition of an inhaled longacting beta 2 agonist (LABA) is now the preferred treatment option [1][2][3][4][5][6]. Addition of a LABA to ICS has been reported to reduce the rate of exacerbations requiring oral corticosteroids, improve lung function, improve symptoms and decrease the use of rescue SABAs, compared with the same dose of ICS alone. If asthma control is not achieved with low dose ICS/ LABA, guidelines recommend that the ICS dose should be increased; therefore it is important to have more than one strength of ICS/LABA available.
The ICS/LABA combination fluticasone furoate (FF)/ vilanterol (VI) is a once-daily asthma treatment, and efficacy over a full 24-h period in asthma has been demonstrated for both FF [7] and VI [8]. In asthma, the optimal doses of oncedaily FF are 100 and 200 mcg [9][10][11][12], and the optimal dose of VI is 25 mcg [13]. Despite the widespread availability of medications for asthma, suboptimal control remains an issue, which may be linked to poor adherence to therapy [14]. Oncedaily regimens have the potential to improve adherence compared with twice-daily regimens [15][16][17]. Good adherence to asthma medication can improve asthma control and contribute to better quality of life for patients, with, for example, fewer unscheduled visits to physicians or fewer missed work/school days [18].
The primary objective of this study was to examine the efficacy and safety of once-daily FF/VI 100/25 mcg versus once-daily FF 100 mcg, over 12 weeks, in patients !12 years of age with moderate-to-severe persistent asthma. The secondary objective was to descriptively assess the relative efficacy of once-daily FF/VI 100/25 mcg versus once-daily FF/VI 200/25 mcg. Preliminary results have been presented in abstract form [19].

Patients
The study was approved by local ethics review committees and conducted in accordance with applicable regulatory requirements, the Declaration of Helsinki [20], and Good Clinical Practice guidelines [21]. All patients (or guardians for patients518 years) gave written informed consent prior to any study procedures being performed.
Key inclusion criteria were age !12 years and moderateto-severe asthma treated with an ICS ± LABA for !12 weeks; a dose that was equivalent to twice-daily fluticasone propionate (FP) 4250 mcg or twice-daily FP/salmeterol 250/ 50 mcg that was stable for !4 weeks. Patients had to have a pre-bronchodilator forced expiratory volume in 1 s (FEV 1 ) 40-80% of predicted normal and FEV 1 reversibility of !12% and !200 mL following salbutamol/albuterol at randomization (historical reversibility, i.e. documented evidence of reversibility within the previous 6 months, was accepted at screening). Key exclusion criteria were: history of lifethreatening asthma within 5 years; any asthma exacerbation requiring oral corticosteroids within 12 weeks or resulting in hospitalization within 6 months; concurrent respiratory disease; use of tobacco products within 3 months or historical use of !10 pack-years (calculated by the number of packs of cigarettes smoked per day multiplied by the number of years, e.g. 10 pack-years is equivalent to smoking 1 pack [20 cigarettes] per day for 10 years).
At screening, patients on ICS continued taking their current daily dose and those using ICS/LABA switched to their current daily dose of ICS alone (LABAs were not permitted from screening onwards) for the 4-week run-in period. All patients replaced their current SABA with salbutamol, which was to be withheld for !6 h prior to study visits. At randomization, eligible patients had to have reported (via an electronic diary) asthma symptoms (a score of !3 on the combined day-and night-time asthma symptom scale) and/or daily salbutamol use on !4 of the last 7 days of the run-in period. Compliance with ICS during the run-in period was reported by the patient via an electronic diary.
Randomization was carried out centrally by the sponsor using a validated computer system (RandAll; GSK, Brentford, UK) and registration and medication ordering system (RAMOS; GSK).
Patients were stratified by asthma severity (severe: baseline FEV 1 65%; or moderate: baseline FEV 1 465% of predicted normal) and randomized to one of three groups in a 1:1:1 ratio: FF 100 mcg (emitted dose of 92 mcg); or FF/VI 100/25 mcg (emitted dose of 92/22 mcg); or FF/VI 200/ 25 mcg (emitted dose of 184/22 mcg). Each medication was to be taken once-daily in the evening for 12 weeks. FF and FF/ VI were administered via the ELLIPTA Õ dry powder inhaler. 1 The inhaler dose counters were reviewed at each visit to assess treatment compliance.
If any patients met the following criteria, indicating lack of efficacy, they were withdrawn from the study: FEV 1 measurement 580% of pre-bronchodilator FEV 1 at randomization; in the 7 days preceding any visit, !4 days in which the peak expiratory flow (PEF) was 580% of the mean morning PEF measured in the week prior to randomization; in the 7 days preceding any visit, !3 days in which !12 inhalations/day of salbutamol/albuterol were used; severe exacerbation, defined as a deterioration of asthma requiring the use of systemic corticosteroids for !3 days, or an in-patient hospitalization/ emergency department visit due to asthma that required systemic corticosteroids; or clinical asthma worsening, which in the opinion of the investigator, required additional asthma treatment other than study medication or study-supplied salbutamol/albuterol. At the randomization visit, patients were trained by the investigator (or a delegate) in the correct use of the inhaler. Placebo inhalers were used as demonstration inhalers. Following the demonstration, the patient's competence with a demonstration inhaler was assessed. If the patient did not use the inhaler correctly, further instructions were given before assessing competence again. If the patient was still unable to use the inhaler correctly after three demonstrations, the patient was considered ineligible to enter the study. The correct use of the inhaler was reassessed at the second and fourth weeks of treatment using the demonstration inhaler. If the patient did not perform the correct technique, the procedure was demonstrated again.

Outcome measurements
The primary end point was the mean change in the 0-24-h post-dose weighted mean (wm) serial FEV 1 from baseline to week 12. The 0-24-h post-dose wmFEV 1 represents the average FEV 1 from a series of measurements taken over a 24h period (5, 15 and 30 min, and 1, 2, 3, 4, 5, 12, 16, 20, 23 and 24-h post-dose) and gives an indication of lung function over the 24-h period.
Secondary efficacy end points that were statistically powered to detect predefined differences between treatments were: change from baseline in (a) trough FEV 1 at week 12 and (b) the proportion of rescue [medication]-free 24-h periods as a percentage of the overall study period during the 12 weeks.
Remaining secondary end points were: (a) change from baseline in the proportion (%) of symptom-free 24-h periods during the 12 weeks; (b) change from baseline in daily morning and evening PEF, averaged over the 12 weeks. Other end points included: the 12-h post-bronchodilator FEV 1 at week 12; the number of withdrawals due to lack of efficacy during the 12-week period and change from baseline in the 1 ELLIPTA Õ is a trademark of the GSK group of companies. Asthma Quality of Life Questionnaire (AQLQ12+) and the Asthma Control TestÔ (ACT) scores at week 12.

Safety evaluations
Safety and tolerability assessments included recording of adverse events (AEs), serious AEs (SAEs), AEs of special interest (class effects of ICS or LABA) and severe asthma exacerbations requiring treatment with systemic/oral corticosteroids for !3 days or an in-patient hospitalization or an emergency department visit due to asthma that required systemic corticosteroids (the latter were not recorded as AEs unless considered to be an SAE). AEs were coded according to the Medical Dictionary for Regulatory Activities.

Statistical analyses
There is no established minimal important difference (MID) for wmFEV 1 in patients with asthma; an assumed treatment difference of 135 mL was based on the largest difference (136 mL) observed in previous studies for the comparison of FF/VI 200/25 mcg versus FF 200 mcg at week 24 [22]. Assuming a true population effect of 135 mL, a sample size of 290 evaluable patients per group would have 97% power in wmFEV 1 over 24 h post-dose.
For the powered secondary end points, 316 evaluable patients per group would have 95% power, assuming a true population effect of 120 mL, to detect a treatment difference in change from baseline in trough FEV 1 , and 499% power, assuming a true population effect of 15%, for the treatment difference in change from baseline in the proportion of rescue-free 24-h periods. The overall power of the study to detect treatment differences between FF/VI 100/25 mcg and FF 100 mcg in the primary and powered secondary end points was 94%. This study was not powered to compare FF/VI 100/ 25 mcg with FF/VI 200/25 mcg.
The intent-to-treat (ITT) population included all randomized patients who received at least one dose of study medication. The per-protocol (PP) population included all ITT patients without any full protocol deviations. The primary and powered secondary end points were analyzed using an analysis of covariance (ANCOVA) model (covariates were baseline, region, sex, age and treatment group).
A step-down, closed-testing procedure was used to account for multiplicity across treatment comparisons and key end points. If the primary end point achieved significance, the secondary end points were tested according to the following priority: trough FEV 1 , rescue-free 24-h periods, symptomfree 24-h periods, morning PEF and evening PEF. A p value50.05 was required to permit inference of significance for the next comparison in the hierarchy. If significance was achieved for each of the secondary efficacy end points, then all other efficacy end points were tested without further multiplicity adjustment.

Study population
Of 2019 patients screened, 1039 were randomized and included in the ITT population, and 931 patients completed the study (Figure 1). The most common reason for withdrawal post-randomization was lack of efficacy. At baseline, patient demographics and characteristics were well balanced across groups, and prior to study entry, the majority of patients (64%) were using ICS/LABA (Table 1). Baseline lung function according to asthma severity is shown in online Supplementary Table S1. Eleven patients used concomitant asthma medication (prescribed or over the counter) during the study, with prednisolone (used by four patients in the FF 100 mcg group and one patient in the FF/VI 100/25 mcg group) being the most common. Non-asthma medication was used by 58-61% of patients pre-study and 65-69% during the study. In all treatment groups, median exposure to study treatment was 84 days and mean overall compliance was approximately 99%.

Efficacy
Primary end point At week 12, there was a mean improvement from baseline in wmFEV 1 0-24 h of 0.366 L with FF 100 mcg and 0.474 L with FF/VI 100/25 mcg; the treatment difference for FF/VI 100/25 mcg versus FF 100 mcg (0.108 L) was statistically significant (p50.001; Table 2 and Figure 2A). Results in the PP population supported the ITT population ( Figure 2A). A larger change from baseline in serial wmFEV 1 occurred in patients with more severe asthma (FEV 1 65%) at baseline versus those with FEV 1 465% (Supplementary Table S2 and Supplementary Figure S1), although for FF/VI 100/25 mcg versus FF 100 mcg, the treatment difference was greater in the FEV 1 465% stratum. The adjusted mean change from baseline of the serial FEV 1 measurements at week 12 showed a relatively flat profile over the 24-h duration, suggesting that respective efficacy levels were maintained during this time for each of the three once-daily treatments ( Figure 3).

Powered secondary end points
At week 12, change from baseline in trough FEV 1 was 0.365 L for FF 100 mcg and 0.441 L for FF/VI 100/25 mcg; the treatment difference for FF/VI 100/25 mcg versus FF 100 mcg of 0.077 L was statistically significant (p ¼ 0.014; Table 2 and Figure 2). The increased trough FEV 1 benefit of FF/VI 100/25 mcg versus FF 100 mcg occurred by week 2 and was sustained ( Figure 4A). A greater change from baseline in trough FEV 1 at week 12 was seen in patients with FEV 1 65% at baseline, but the treatment difference for FF/VI 100/25 mcg versus FF 100 mcg was higher in the FEV 1 465% than in the FEV 1 65% stratum (Supplementary  Table S2).
Over 12 weeks, the proportion of rescue-free 24-h periods increased from baseline in all groups (Table 2 and Figure 4B). The treatment difference was 12.2% (p50.001) between FF/ VI 100/25 mcg and FF 100 mcg (Table 2 and Figure 2A), equivalent to 0.9 additional rescue-free days per week.

Secondary and other end points
The proportion of symptom-free 24-h periods increased from baseline over 12 weeks in all groups (Supplementary Table  S3 and Figure 2A), which equated to 0.5 additional symptomfree 24-h periods per week.
Morning PEF was increased from baseline in all groups (Supplementary Table S3 and Supplementary Figure S2A), and there was a statistically significant (p50.001) treatment difference (25.2 L/min) between FF/VI 100/25 mcg and FF 100 mcg (Supplementary Table S3 and Figure 2A). Similarly, evening PEF was increased from baseline (Supplementary Table S3 and Supplementary Figure S2B); the treatment difference for FF/VI 100/25 mcg versus FF 100 mcg of 24.2 L/ min was statistically significant (p50.001; Supplementary Table S3 and Figure 2A).
Data for other end points are shown in Supplementary Table S3 and Figure 2(A). The treatment difference for FF/VI 100/25 mcg versus FF 100 mcg for the change from baseline in 12 h post-dose FEV 1 at week 12 (0.124 L) was statistically significant (p50.001). Although increases from baseline in AQLQ12+ score at week 12 in all groups were greater than the MID of 0.5 [23,24], the treatment difference of 0.08 (p ¼ 0.303) between the FF/VI 100/25 mcg and FF 100 mcg groups was smaller than the MID. Significantly fewer patients withdrew due to lack of efficacy in the FF/VI 100/25 mcg group versus FF 100 mcg (p ¼ 0.003); this is supported by the cumulative incidence of time to withdrawal (Supplementary Figure S3). A statistically significant (p ¼ 0.002) treatment difference for FF/VI 100/25 mcg versus FF 100 mcg of 0.9 in ACT score was observed at week 12. The proportion of patients with well-controlled asthma (ACT score ! 20) at week 12 was 40% (FF 100 mcg) and 46% (FF/VI 100/25 mcg  ITT, intent-to-treat; PP, per-protocol; VI, vilanterol. a One patient had enrolled at each of two US sites which was discovered after randomization and this patient was withdrawn from both sites.

Discussion
In this study, once-daily FF/VI 100/25 mcg significantly improved wmFEV 1 0-24 h, an important measure of lung function over the 24-h dosing interval, versus FF 100 mcg in patients with asthma previously uncontrolled on mid-to-high dose ICS or mid-dose ICS/LABA. This was supported by significant improvements in trough FEV 1 and the proportion of 24-h periods that patients did not have to use rescue medication and the proportion of 24-h periods that patients did not have symptoms.
Previous studies comparing FF/VI 100/25 mcg and FF 100 mcg have shown mixed results for FEV 1 end points. In a 12-week study, Bleecker et al. were not able to show statistically significant treatment differences in wm and trough FEV 1 end points for once-daily FF/VI 100/25 mcg over FF 100 mcg [16] in patients uncontrolled on low-to-middose ICS or low-dose ICS/LABA; in our study, however, patients using low-dose ICS ± LABA were not included. In a study by Bateman et al., which recruited patients receiving !200 mcg FP daily or equivalent, or 200/100-500/100 mcg FP/salmeterol daily or equivalent, improvements from baseline in trough FEV 1 were significant for FF/VI 100/ 25 mcg versus FF 100 mcg [25]. Our study provides further information on the contribution of VI to the efficacy of FF/VI 100/25 mcg compared with FF 100 mcg alone, particularly with respect to the serial measurement of wmFEV 1 0-24 h (which was not assessed by Bateman et al.). It is worth noting that the treatment difference in wmFEV 1 0-24 h for FF/VI 100/25 mcg versus FF 100 mcg in the Bleecker et al. study (116 mL) was slightly larger than in the present study (108 mL) [12]. Thus, a more likely explanation for the discrepant results in wmFEV 1 is the difference in statistical power between the studies; fewer patients in the Bleecker et al. study underwent serial FEV 1 evaluations than in our study.
For wmFEV 1 and trough FEV 1 , the change from baseline with FF 100 mcg and FF/VI 100/25 mcg was larger for patients with baseline FEV 1 65% predicted compared with those with FEV 1 465% predicted. These observations indicate that patients with poorer lung function at baseline may have a greater potential to improve lung function in response to anti-inflammatory treatments. FF/VI 100/25 mcg significantly improved all secondary and other end points (except change from baseline in AQLQ at week 12) versus FF 100 mcg. These findings are consistent with previous studies comparing FF/VI with FF [12,22] and with meta-analyses comparing ICS/LABA with ICS only in patients with persistent asthma [6,26]. Improvements in the proportion of rescue-free 24-h periods and symptom-free 24-h periods seen with FF/VI 100/25 mcg versus FF 100 mcg were clinically meaningful, being within the MIDs previously identified (8-16% and 8-15%, respectively) [27]. Although the change from baseline values for each treatment exceeded the MIDs (0.5 [23,24] and 3 units [28], respectively), indicating that patients had improved quality of life and asthma control with each of the once-daily treatments, the treatment differences between FF/VI 100/25 mcg and FF 100 mcg in AQLQ and ACT scores failed to reach the MIDs [21, 22,26]. However, the odds of being well-controlled at week 12 (ACT score ! 20) were 42% greater for patients treated with FF/VI 100/25 mcg compared with FF 100 mcg. A previous study that recruited patients with a history of severe exacerbations [25] reported that FF/VI 100/25 mcg was associated with greater improvements in Asthma Control Questionnaire 7 (ACQ7) than FF 100 mcg (p50.001 at all time points) with the odds of a patient having well-controlled asthma (ACQ7 score ! 0.75) being greater for FF/ VI compared with FF. At the study endpoint, patients receiving FF/VI were 50% more likely to have their asthma well-controlled than those on FF alone [25].
A secondary objective of this study was to descriptively assess the relative efficacy of once-daily FF/VI 200/25 mcg versus FF/VI 100/25 mcg. Small numerical improvements were seen with FF/VI 200/25 mcg versus FF/VI 100/25 mcg for all end points. It can be difficult to show a dose-response to ICS, especially in the presence of a LABA; however, the odds of a patient being well-controlled (ACT score ! 20) were 55% higher with FF/VI 200/25 mcg compared with FF/VI 100/25 mcg. These results are supported by those of a study of different doses of FF monotherapy, which showed numerical increases in lung function end points for once-daily FF 200 mcg compared with once-daily FF 100 mcg [29]. In that study, the odds of being well-controlled (ACT score ! 20) were 42% greater with FF 200 mcg compared with FF 100 mcg. In the present study, for wmFEV 1 and trough FEV 1 , the change from baseline with FF/VI 200/25 mcg versus FF/ VI 100/25 mcg was greater for patients with FEV 1 465% than those with FEV 1 65% at baseline (62 mL versus À1 mL and 57 mL versus À12 mL, respectively). These observations suggest that ICS dose-response may be dependent on the level of airway inflammation, which may be reflected by the severity of airflow limitation.
Overall, no new safety concerns were raised with any treatments; all options were well-tolerated and represent valid treatment choices for patients. The AEs with FF/VI 100/ 25 mcg and FF/VI 200/25 mcg were consistent with those reported in a 12-month safety study [30]. The incidences of AEs expected with ICS (e.g. local steroid effects, respiratory tract infections, bone disorders [1,31]) or LABA (e.g. glucose increase, cardiovascular effects [1]) were low across all groups. As expected, local steroid effects were observed in all groups, although the incidence was low and similar across all treatment groups. There was no evidence of an increased risk of cardiovascular effects due to the addition of VI to the regimen. Although some patients experienced severe asthma exacerbations requiring oral corticosteroids during the study, the incidence was less than that recorded in the year prior to the study.
The study strengths include enrollment of suitable patients with persistent moderate-to-severe asthma; i.e. patients with uncontrolled asthma and requiring Step 3 treatment or above, as per the Global Initiative for Asthma (GINA) guidelines [1]. Compliance with study medication in all groups was very high (&99%). Moreover, this study is one of the first to assess two doses of the same ICS (i.e. FF) in the corresponding ICS/ LABA combination. Potential limitations of our study include the lack of a placebo arm, the relatively short duration and the fact that the comparison of FF/VI 200/25 mcg versus FF/VI 100/25 mcg was not powered and could only be interpreted descriptively. There were only a few patients on high-dose ICS pre-study, so an evaluation of the potential dose response between the two doses of FF/VI in such patients was not feasible. Additionally, the proportion of patients with childhood-onset asthma or allergic rhinitis, neither taking intranasal corticosteroids was not recorded, nor were any environmental controls specified in the current study. Allergic rhinitis and environmental factors (e.g. second-hand smoke) are known to affect asthma control [32,33] and could potentially confound the interpretation of our results.

Conclusions
This study demonstrated that, in patients with moderate-tosevere persistent asthma, once-daily FF/VI 100/25 mcg was significantly more effective in improving lung function and symptoms than once-daily FF 100 mcg, over 12 weeks. Small numerical improvements occurred with FF/VI 200/25 mcg versus FF/VI 100/25 mcg. AEs were in line with the class effects expected for ICS and LABA treatment. The incidence of drug-related AEs was low with each of the studied treatment options. Very few drug-related AEs led to withdrawal of treatment, suggesting that FF/VI 100/25 mcg, FF/VI 200/25 mcg and FF 100 mcg are all valid treatment options for moderate-to-severe persistent asthma, providing further choice for patients. The study was sponsored by GSK (study number HZA116863; www.clinicaltrials.gov registration number NCT01686633). Employees of the sponsor were involved in the conception, design and conduct of the study, and in data collection and analysis. All authors, including authors employed by the sponsor, participated in the development of the manuscript, and had access to the data from the study. The decision to submit for publication was that of the authors alone. All data are n (%). AE, adverse event; FF, fluticasone furoate; ITT, intent-to-treat; OD, once daily; VI, vilanterol.