18 November 2009,
Two one-year RCTs (AURA and HERMES) showed that roflumilast modestly improved lung function in patients with severe COPD who were symptomatic and not using an inhaled steroid. A reduction in moderate-to-severe exacerbations was also reported (NNT=5 over one year). Two six-month RCTs (EOS and HELIOS) found similar benefits in lung function when roflumilast was added to salmeterol or tiotropium in patients with moderate-to-severe COPD. In all of the studies, more patients discontinued treatment with roflumilast because of adverse effects (NNH=35); the most common of which were GI side effects (e.g. nausea, diarrhoea), weight loss and headache. Further studies are required to establish whether roflumilast offers any meaningful advantages as an alternative, or in addition, to an inhaled steroid.
Level of evidence
Level 2 (limited patient-oriented evidence) according to the SORT criteria
The place of roflumilast in the treatment of patients with COPD is uncertain. The four studies reviewed here provide insufficient evidence to justify its use ahead or instead of inhaled corticosteroids (ICSs), for those people who remain symptomatic or suffer exacerbations on optimal bronchodilator therapy (long-acting beta-agonists and/or tiotropium). Any possible benefits of roflumilast have to be balanced against the possibility of adverse effects, most commonly diarrhoea, nausea, headache, and weight loss.
Roflumilast has been submitted for an EU license, but no NICE guidance is planned currently. Therefore, local decision making bodies on medicines are advised to engage with stakeholders to identify those patients for whom the drug may be appropriate and agree a protocol for use if a license is granted and the product is launched.
What is the background to this?
Roflumilast (Daxas®, Nycomed) is a once-daily, oral phosphodiesterase-4 inhibitor, which has anti-inflammatory properties, being developed for the treatment of COPD. A marketing authorisation application was filed for the EU in May 2009, and if a licence is granted, it is likely to be launched in 2010. This licence application is based on two pivotal one-year, placebo-controlled randomised clinical trials (AURA and HERMES), and supported by two six-month, placebo-controlled trials (EOS and HELIOS), in which it was used with a long-acting bronchodilator (salmeterol or tiotropium). AURA and HERMES compared the effect of roflumilast 500micrograms daily with placebo in patients with severe COPD, bronchitic symptoms and a history of exacerbations. They had identical protocols, but used two geographically different populations. The main difference between the two populations was that HERMES included a greater proportion of people of Asian ethnic origin than AURA (23% vs. <1%). EOS and HELIOS evaluated the addition of roflumilast 500micrograms daily to either salmeterol or tiotropium, respectively, in patents with moderate-to-severe COPD. However, the patients in these latter trials did not necessarily have to have a history of exacerbations.
The current NICE clinical guideline for COPD from 2004 recommends that people with moderate-to-severe COPD, who are still symptomatic despite use of a long-acting bronchodilator should be given a four-week trial of an ICS in addition to a long-acting bronchodilator. Furthermore, the ICS should be discontinued if there is no benefit after four weeks (if still symptomatic then theophylline may be considered in addition). An ICS is also recommended where the forced expiratory volume in one second (FEV1) is £50% predicted and the patient suffers two or more exacerbations in a 12-month period; normally this would be in addition to a long-acting bronchodilator. Note that an update to the NICE clinical guideline is due to be published in June 2010; and there are proposed changes to the currently recommended treatment regimen in the draft consultation guideline.
Further information about COPD can be found on the respiratory tract floor of NPC.
What do these studies claim?
Similar results for AURA and HERMES for the primary endpoints (change in prebronchodilator FEV1 and rates of exacerbations) were claimed. Pooled results indicated that roflumilast produced a modest improvement in pre-bronchodilator FEV1 compared with placebo of 48mL (95% confidence interval [CI] 35mL to 62mL; P<0.0001); individual results for AURA and HERMES were 39mL and 58mL, respectively. There was a mean reduction in the rate of moderate (i.e. requiring oral corticosteroids or parenteral corticosteroids) or severe exacerbations (i.e. resulting in hospital admission or death) from 1.37 per year to 1.14 per year (rate ratio [RR] 0.83; 95%CI 0.75 to 0.92; P=0.0003) with roflumilast — a reduction of 0.23 exacerbations per year (number needed to treat (NNT) 5 to prevent one exacerbation in a year).
Results of EOS and HELIOS studies are difficult to compare with eachother, because patients in the HELIOS study were more symptomatic than in EOS, and used more as-needed medication. Compared with placebo, roflumilast improved mean prebronchodilator FEV1 (the primary endpoint) by 49mL (95%CI 27 to 71; P<0.0001) when added to salmeterol and 80mL (95%CI 51 to 110; P<0.0001) when added to tiotropium.
Discontinuations due to adverse events were more common with roflumilast than placebo (significant in AURA, HERMES and EOS). Gastrointestinal side-effects (e.g. diarrhoea, nausea), weight loss, or headache were the most common side effects associated with roflumilast.
How does this relate to other studies?
Two one-year, RCTs of roflumilast versus placebo in COPD have been published previously:
- A study of 1,411 patients with moderate-to-severe COPD considered two doses (250 and 500micrograms daily) of roflumilast. It found a statistically significant improvement in postbronchodilator FEV1 with roflumilast compared with placebo (97mL with 500micrograms daily). There was no significant effect on health-related quality of life. Although there was a significant difference in favour of roflumilast with regard to the rate of exacerbation, this was predominantly due to a reduction in mild exacerbations; the rates of moderate or severe exacerbations were similar between groups.
- A study of 1,513 patients with severe, stable COPD demonstrated a statistically significant mean improvement in postbronchodilator FEV1 of 39mL for roflumilast compared with placebo. However no differences were found with regard to the rate of exacerbations or health status. Post-hoc analysis suggested one of the hypotheses tested in the AURA and HERMES studies, i.e. that roflumilast may reduce exacerbations in specific subsets of patients with severe COPD.
The results of AURA and HERMES suggest that roflumilast modestly improves lung function compared with placebo in patients with severe COPD who have bronchitic symptom, and who have a history of exacerbations. They also demonstrate a reduction in the rate of exacerbations — about five patients would be treated with roflumilast for a year, for one of them to have one fewer moderate or severe exacerbation during that period. Although some patients may obtain benefit in this regard, this potential benefit has to be put into context by considering adverse effects of treatment and the limitations of the study.
Although these were well controlled studies, they do not inform us about whether roflumilast provides any benefits when used instead of, or in addition to, the current NICE recommended treatment regimens. Many of the patients in these studies were already receiving an ICS and long-acting bronchodilators, which were discontinued, although continued use of LABAs was allowed during the studies. Questions remain, therefore, of whether or not roflumilast offers any significant benefit when used instead of inhaled corticosteroids, or when used in addition to inhaled corticosteroids and/or long-acting bronchodilators. Until these questions are answered by further clinical studies it is difficult to see where roflumilast sits in the routine treatment of COPD.
The question of whether there are any benefits when used in addition to long-acting bronchodilators is to some extent answered by the results of the EOS and HELIOS studies. A modest statistically significant improvement in lung function was identified in the patients in both studies of patients who had less severe COPD (mostly moderate). Exacerbation rates were low during both studies, which were underpowered to test differences between treatments in this respect. As in AURA and HERMES, an ICS was not allowed, and they provide no information with which to compare the benefits of roflumilast with an ICS or indeed an additional different class bronchodilator (i.e. when salmeterol or formoterol is used together with tiotropium).
Any potential benefits of roflumilast have to be balanced against the risk of adverse effects. All studies demonstrated increased incidence of adverse effects compared with placebo. Gastrointestinal side effects, headache and weight loss were reported as being significantly more common with roflumilast than with placebo in most or all of the studies. Although these side effects may diminish over time, they may not be tolerated by some patients. In AURA and HERMES more patients discontinued treatment due to adverse events (14% vs. 11%, number needed to harm [NNH] 35).
It is notable that any improvements in AURA and HERMES in overall health utility over placebo were small and not statistically significant.
1) AURA (M2-124) and HERMES (M2-125)
Calverley PMA, Rabe KF, Goehring U-M, et al. Roflumilast in symptomatic chronic obstructive pulmonary disease: two randomised clinical trials. Lancet, 2009; 374: 685–94
Design: Both studies were multicentre, international, one-year, double-blind, placebo-controlled, randomised clinical trials with identical designs in two different geographical populations. Analysis was by intention to treat. Allocation was concealed.
Patients: Outpatients with severe COPD with chronic cough and sputum production: post-bronchodilator FEV1 £50% predicted with at least one moderate or severe exacerbation in the previous year; older than 40 years of age (average 64 years both studies); 71% men in AURA, 81% men HERMES. In HERMES the proportion of people of Asian origin was higher (23% vs. <1%), mean baseline pre-bronchodilator and post-bronchodilator FEV1 were lower (0.97L vs. 1.07L, and 1.06L vs. 1.16L), there were fewer current smokers (35% vs. 48%) and more patients had very severe COPD (33% vs. 25%). Overall 42% had used ICS previously and 51% used LABAs during the trials.
Interventions/Comparison: After a 4-week run-in on placebo, patients received either roflumilast 500micrograms daily or placebo for 52 weeks. Short-acting beta-agonists, as needed, and regular short-acting and long-acting beta-agonists could be continued throughout the study at stable doses. ICSs and tiotropium were not allowed during the study.
Outcomes and Results:
The primary endpoints were the change in pre-bronchodilator FEV1 during treatments and the rate of COPD exacerbations. Pooled results indicated that roflumilast produced a modest improvement in pre-bronchodilator FEV1 compared with placebo of 48mL (95%CI 35mL to 62mL; P<0.0001); When analysed individually, changes in this primary outcome for AURA and HERMES were also significantly different from roflumilast compared with placebo (39mL and 58mL, respectively; both P<0.001). Pooled results indicated a mean reduction in the rate of moderate (i.e. requiring oral corticosteroids or parenteral corticosteroids) or severe exacerbations (i.e. resulting in hospital admission or death) from 1.37 per year to 1.14 per year (RR 0.83, 95%CI 0.75 to 0.92; P=0.0003) with roflumilast. The mean reduction of 0.23 exacerbations per year (NNT to prevent one exacerbation in a year is 5); reductions in the rates of exacerbations for individual studies were 0.19 and 0.28 (P=0.028 and P=0.0035, respectively). Pooled results (and results from individual studies) did not show a significant difference in the rate of severe exacerbations, although the study was underpowered to evaluate this outcome. Roflumilast did not reduce the median time to first exacerbation (moderate or severe) significantly in the individual studies; pooling of results suggested a reduction of only 9 days on average (71 vs. 80 days, P=0.019).
There was a small but clinically insignificant change (i.e. less than1point) in the transition dyspnoea index, no significant change in time to death, or change in Euroqol-5 dimension score (a measure of health utility).
There were slightly more withdrawals in the roflumilast groups compared with the placebo groups (AURA, 35% vs. 31%; HERMES 32% vs. 31%). Discontinuations associated with adverse events were more common in the pooled roflumilast groups (14% vs. 11% overall [NNH 35]; 8% vs. 3% in the first 12 weeks. With the exception of COPD, the most frequent adverse events leading to discontinuation were diarrhoea, nausea and headache. Significantly more adverse events were identified in both studies for roflumilast with regard to diarrhoea, weight loss (about 2 kg after a year) and decreased appetite. Nausea and headache was only reported significantly more frequently in AURA.
2) EOS (M2-127) and HELIOS (M2-128)
Fabbri LM, Calverley PMA, Izquierdo-Alonso JL, et al. Roflumilast in moderate-to-severe chronic obstructive pulmonary disease treated with longacting bronchodilators: two randomised clinical trials. Lancet 2009; 374:695-703.
Both studies were multicentre, international, placebo-controlled, 24-week, double-blind, placebo-controlled trials. Analysis was by intention to treat. Allocation was concealed.
Patients: Outpatients with moderate-to-severe COPD: post-bronchodilator FEV1 40–70% predicted; older than 40 years of age. All patients in HELIOS had chronic cough and sputum production (78% in EOS) and were required to use more than 28 puffs per week of as-needed short-acting beta agonists. In EOS and HELIOS, respectively, mean ages were 65 and 64 years and the proportions of men in the studies were 66% and 72%; mean baseline pre-bronchodilator FEV1s were 1.42L and 1.48L. About a third of patients in both studies had severe COPD.
Interventions/comparison: After a four week-run in on placebo, patients were treated with either roflumilast 500micrograms daily or placebo in addition to either salmeterol (EOS) or tiotropium (HELIOS) for 24 weeks. No ICSs, inhaled short-acting bronchodilators, or other respiratory drugs were allowed during the study (although rescue medication was used during the study).
Outcomes and Results:
The primary endpoint was the change in pre-bronchodilator FEV1. Compared with placebo, roflumilast improved mean prebronchodilator FEV1 by 49mL (95%CI 27 to 71; P<0.0001) when added to salmeterol and 80mL (95%CI 51 to 110, P<0.0001) when added to tiotropium. Statistically significant improvements in other measurements of lung function were also identified. The study was not powered to compare the rates of exacerbations or other endpoints, and there were no significant improvements in the rate of mild, moderate or severe exacerbations in either study. Although there were statistically (but not necessarily clinically) significant improvements compared with placebo for roflumilast when added to tiotropium in HELIOS in TDI focal score, shortness of breath questionnaire reduction, and reduction in the use of rescue medication, no significant differences were identified for these secondary endpoints in EOS.
Withdrawals from the study were higher with roflumilast in both studies (23% vs. 18% EOS; 17% vs. 10% HELIOS). Adverse events related to treatment were higher with roflumilast in both studies (EOS, 18% vs. 3%; HELIOS 12% vs. 2%). Diarrhoea, nausea and weight loss (about 2kg during the studies) were significantly more common with roflumilast than placebo in both studies.
All four studies were either sponsored by Nycomed or Altana, the previous licensee for the product.
Make sure you are signed up to NPC Email updates — the free email alerting system that keeps you up to date with the NPC news and outputs relevant to you.