Anti-inflammatory duration of action of fluticasone furoate/vilanterol trifenatate in asthma: a cross-over randomised controlled trial

Combination inhaled corticosteroid (ICS) / long-acting beta-agonist (LABA) therapy forms the cornerstone of guideline recommended management of moderate to severe asthma [1] and in some countries ICS/LABA combination inhalers are the most commonly prescribed ICS containing medication [2]. Most ICS/LABA medications require twice-daily dosing, however adherence to ICS-containing therapy is low and changing inhaled therapy to a once-daily regimen is suggested by international guidelines as a strategy for improving adherence [1, 3]. These findings have led to the development of ICS and LABA medications with a prolonged duration of action, allowing once-daily dosing.

Fluticasone furoate / Vilanterol trifenatate (FF/VI) 100/25 mcg is an ICS/LABA combination licensed for the once-daily treatment of asthma and chronic obstructive pulmonary disease [4]. In people with asthma, after a single dose of FF/VI, bronchodilation was maintained for 72 h post dose with clinically significant bronchodilation persisting for at least 48 h [5]. However, the onset and duration of anti-inflammatory action of FF/VI has not been studied and since LABA monotherapy without concomitant ICS therapy is associated with increased mortality in asthma [6], it is important to determine whether the anti-inflammatory duration of action of FF/VI is of at least the same duration as the bronchodilator activity. Although this is of limited relevance for chronic therapy in treatment adherent patients it may be important for patients who miss one or more doses. The primary objective of this study was to characterise the duration of anti-inflammatory activity of FF/VI administered to patients with asthma. We hypothesised that FF would exhibit a prolonged duration of anti-inflammatory action, as the FF component is an ICS with higher glucocorticoid receptor affinity [7] and more prolonged lung retention [8], compared to other ICS.

The study was conducted between April 2016 and February 2017, during which 32 potential participants were screened. Four participants were excluded on the grounds of a screening FeNO ≤40 ppb and 28 participants were enrolled. One participant withdrew after screening but prior to taking any study medication. 27 participants successfully completed both study periods and were included in the final analysis (Additional file 1: Figure S1).

Baseline characteristics of the 27 randomised participants are shown in Table 1. Participants had mild to moderate airflow obstruction at baseline, mean (standard deviation) FEV₁ percent predicted 87.7 (10.6). Median (range) baseline FeNO and blood eosinophils were 87 ppb (42 to 212) and 0.33 × 10^9/L (0.18 to 1.18) respectively.

The time-course of FeNO during the 14-day treatment and 21-day washout periods is shown in Table 2 and Fig. 2. Summary data by treatment period are provided in Additional file 1: Table S2. Compared with placebo, FF/VI 100/25mcg treatment decreased the FeNO by the third day following onset of treatment, change from baseline ratio FF/VI versus placebo 0.72 (95% CI 0.61 to 0.86). The maximum reduction in FeNO occurred on day 14, change from baseline ratio FF/VI versus placebo 0.32 (0.27 to 0.37).

Following cessation of FF/VI 100/25mcg treatment the FeNO gradually increased towards the values following placebo, remaining suppressed until approximately day 18, ratio 0.77 (0.59 to 1.00). By inspection of Fig. 2 it can be seen that the rate of onset of FeNO suppression was greater than the rate of offset.

FF/VI 100/25 mcg was associated with significant bronchodilation as measured by FEV₁ (Table 3, Fig. 3a), adjusted mean FEV₁ change from baseline compared with placebo 0.38 (0.28 to 0.48), 12 h after the last dose on day 14. After cessation of FF/VI treatment, FEV₁ remained significantly different from placebo until the PM measurement on Day 4. The increase in FEV₁ with FF/VI relative to placebo was greater than the minimum perceivable improvement (0.23 L) [10] until day 3 after cessation of treatment, difference 0.18 L (0.08, 0.28). PEF remained significantly greater than placebo until the AM measurement on Day 4 (Additional file 1: Table S1, Fig. 3b).

Blood eosinophil and serum periostin levels throughout the study are shown in Table 4. Baseline blood eosinophils were ≥ 0.27 × 10^9/L for 24 participants (88.9%), but the majority of participants had blood eosinophil levels within the normal range throughout the study. Blood eosinophils were reduced after 14 days treatment with FF/VI 100/25 mcg; percentage change from baseline, FF/VI minus placebo, (95%CI), − 17.9% (− 33.2 to − 2.64). The point estimate for serum periostin level change from baseline was lower after FF/VI treatment, however the confidence interval included zero treatment difference, − 6.3% (− 18.5 to 5.9).

No serious adverse events were reported during the study. Seventeen (63%) subjects reported adverse events during the study (Additional file 1: Table S3). The most frequently reported adverse events in the study were viral upper respiratory tract infection (reported by 10 subjects, 37%); and back pain, musculoskeletal pain, and epistaxis (reported by 2 subjects each, 7%). Only one adverse event, cough with moderate intensity, was reported as potentially drug related. This resolved prior to cessation of study medication.

In this placebo-controlled study of the time-course of anti-inflammatory activity of FF/VI in asthma, treatment with FF/VI 100/25 mcg for 14 days led to the suppression of FeNO for approximately 18 days after withdrawal of treatment. In contrast, following cessation of FF/VI treatment, significant bronchodilation persisted for three to four days. This indicates that FF/VI exhibits a prolonged anti-inflammatory effect consistent with the high glucocorticoid receptor affinity and long lung retention of fluticasone furoate. The anti-inflammatory activity of FF/VI lasts longer than the bronchodilator effect, and therefore unopposed bronchodilator activity is unlikely to occur with FF/VI, a finding relevant to the scenario where patients omit one or more doses.

There are a number of methodological issues pertinent to the interpretation of the study findings. FeNO was used as the primary measure to monitor the anti-inflammatory activity of FF/VI as it is a validated biomarker of Type 2 inflammation in asthma, highly sensitive to ICS therapy, and, unlike alternative biomakers such as sputum or blood eosinophils or serum periostin, can be repeated twice daily in both a home and inpatient setting [11]. To ensure that the anti-inflammatory activity of FF/VI could be detected using FeNO, a screening FeNO > 40 ppb was used as an inclusion criterion. A FeNO value above 40 ppb predicts responsiveness to ICS in patients with non-specific airways disease, and is the recommended cut-off denoting a ‘high’ FeNO in the National Institute for Clinical Excellence (NICE) asthma guidelines [12, 13]. As FeNO is suppressed by ICS treatment and cigarette smoking, participants were required to be steroid naïve and not current smokers at baseline. We advised participants to follow a low nitrate diet throughout the study, and strictly ensured this was the case during their inpatient stay. The fall in FeNO in both groups towards the end of the inpatient period may be due to a combination of enforcement of the low nitrate diet and reduced exposure to allergens and environmental pollutants. A 21-day washout period was employed during which time participants were only allowed to take their SABA, with the aim of ensuring that the effect of FF/VI on FeNO had fully resolved. The prolonged time-course of anti-inflammatory activity seen in this study means that for subjects who received FF/VI in period 1 their pre-dose baseline for period 2 was approximately 11% lower than the period 1 pre-dose value. For subjects who received placebo in period 1 their pre-dose baseline for period 2 was approximately 2% lower than the period 1 pre-dose value (Additional file 1: Table S2). However, change in FeNO was assessed relative to the pre-dose value at the start of each treatment period, and in the statistical model used to estimate the primary endpoint “period” was included as a term and shown not to be a significant factor, confirming that there was no significant treatment carry-over between periods. Therefore any trend for a lower FeNO value at the start of period 2 was small and not a confounding factor in the study or its conclusions.

Participants were required to demonstrate a minimum FEV₁ of at least 60% predicted as a safety criterion, and at least 12% bronchodilator reversibility at baseline in order to confirm the diagnosis of asthma. The requirement for a steroid naïve population who were able to tolerate ICS treatment being withheld for up to 21 days required us to recruit only participants with mild asthma. This, in addition to the exclusion of smokers and patients with FeNO levels < 40 ppb, mean that these results may not be generalizable to a wider asthmatic population. Participants were required to remain as inpatients for five days after cessation of treatment to ensure that their environment was as tightly controlled as possible, to avoid exposures that may influence FeNO or lung function measurements. This also allowed us to measure FEV₁, which is a more sensitive measure of airflow obstruction than PEF and is a recommended core outcome measure in clinical trials [14].

It was not possible to determine the relative contributions of FF and VI to the anti-inflammatory and bronchodilator actions of combination FF/VI. However, based on the known properties of FF and VI demonstrated in the clinical trial program that lead to the registration of FF/VI [15], and previous reports that short-acting beta-agonists and LABAs do not affect FeNO [16, 17], the anti-inflammatory activity is likely to be due to the FF component.

FF/VI reduced FeNO within three days of the onset of treatment, with a progressive reduction to about one third of the baseline value by day 14. This is similar to the time-course of onset of FeNO suppression previously reported for inhaled budesonide, fluticasone propionate (FP), ciclesonide, and combination extra fine beclomethasone dipropionate/formoterol [18‐21]. In contrast the suppression of FeNO persisted for 18 days after cessation of FF/VI treatment, which is appreciably longer than that observed with other ICS. Following treatment with budesonide (100mcg or 400mcg once-daily) [20] for three weeks, the FeNO returned to baseline levels by the end of a one week washout period. Following two weeks of treatment with beclomethasone diproprionate (200mcg/day) the FeNO returned to baseline within one to two weeks in the majority of participants [22]. Similarly, following FP (500mcg twice-daily) [19] the FeNO returned to baseline after two weeks, although no earlier time point measurements were included.

In addition to the changes seen with FeNO, two weeks of treatment with FF/VI reduced the blood eosinophil count by 17%. This finding is in keeping with the limited data available on the effect of ICS on blood eosinophils. In two separate studies in people with moderate asthma, two weeks of FP / salmeterol 100mcg/50mcg twice daily [23] and six weeks of budesonide 800mcg per day [24] were associated with reductions in blood eosinophils of 23 and 26% respectively. The reduction in blood eosinophils with ICS treatment is likely to be associated with a reduction in sputum eosinophils, as blood eosinophils are very strongly correlated with sputum eosinophils [23]. Blood eosinophil levels, although within the normal range for the majority of participants, were ≥ 0.27 × 10^9/L at baseline for 24 (88.9%). This threshold has previously been shown to have a 78% sensitivity and 91% specificity for sputum eosinophilia in participants with mild to moderate asthma [25], so both baseline FeNO and blood eosinophil data are consistent with participants in this study having eosinophilic airway inflammation at baseline. Periostin is associated with a type 2 pattern of inflammation [26], the sensitivity and specificity of the assay used in this study for the prediction of airway eosinophilia has not yet been determined. The point estimate of a 6% fall in periostin is in keeping with prior published data, with 5 and 12% reductions in periostin reported with FP 400mcg daily [27] and budesonide 800mcg daily [28] respectively.

It is difficult to compare the FeNO findings with other ICS as different equivalent daily doses were investigated, which confounds the assessment of responsiveness, as both the onset and offset of ICS-induced reduction in FeNO may be dose dependent [20, 29‐31]. In addition, the baseline FeNO is a determinant of the magnitude and speed of ICS response [20], and the pre-treatment FeNO levels and sampling methodologies were markedly different between studies, with a range of between 6.9 ppb [19] and 122 ppb [32], compared with the baseline value of 87 ppb observed in this study. It is notable that for FF the offset of FeNO suppression was slower than the onset of action. This differs from the pattern previously reported with budesonide where, although the rate of onset and offset of action differed between the 100mcg and 400mcg doses, the rates of onset and offset were similar at each dose [20].

The ICS/LABA combination product studied in this investigation contains two molecules that both contribute to its long duration of action. Compared to other ICS, fluticasone furoate exhibits higher binding affinity and a longer duration of occupancy of the ligand binding domain of the glucocorticoid receptor [7]. The structural features of the FF molecule also confer physicochemical properties that result in higher affinity for respiratory tissue and longer lung retention compared to other ICS. Through in vitro and in vivo studies, FF has been demonstrated to have a longer intracellular duration of action than FP [33], as well as higher potency [33], prolonged intracellular retention [7] and prolonged lung absorption kinetics [7, 8]. Vilanterol has a high β₂ receptor affinity, with rapid onset of action and relatively slow dissociation [34]. The prolonged bronchodilator action of vilanterol may be due primarily to it being a highly lipophilic molecule, leading to the formation of ‘depots’ in the cell membrane resulting in persistence and reassertion at the β₂-receptor [34].

This prolonged anti-inflammatory effect of FF/VI may have implications for poorly adherent patients as it confirms that the anti-inflammatory activity works alongside the bronchodilator activity for the approved once-daily dosing regimen, and also suggests that patients who miss one or more doses of FF/VI will still receive both anti-inflammatory and bronchodilator coverage for some time after the last dose. Finally, it provides reassurance that they will not be exposed to unopposed LABA activity as they are likely to still derive benefits from its anti-inflammatory effect. However, dosing less frequently than once a day has not been studied and FF/VI is only licensed for use as a once-daily medication.

Among adult patients with asthma who were not taking ICS at baseline, treatment with FF/VI at a dose of 100/25mcg once daily for 14 days substantially reduced airway inflammation, as measured by FeNO suppression. After cessation of treatment, FeNO suppression persisted for approximately 18 days, whereas clinically significant bronchodilation persisted for three to four days. These results show that the duration of anti-inflammatory action of FF/VI is longer than has previously been reported for other ICS studied and exceeds the duration of bronchodilation. Therefore, even when adherence to therapy is poor, there is prolonged anti-inflammatory activity and no unopposed LABA activity with this combination ICS/LABA.