Optimizing bronchodilation in the prevention of COPD exacerbations

The natural disease course of chronic obstructive pulmonary disease (COPD) is often punctuated by exacerbations: acute events of symptom worsening associated with significant morbidity and healthcare resource utilization; reduced quality of life; and increased risk of hospitalization and death. The Global Initiative for Chronic Obstructive Lung Disease (GOLD) recommend that patients at risk of exacerbations (GOLD Groups C and D) receive a long-acting muscarinic antagonist (LAMA) or a long-acting β2-agonist (LABA)/LAMA combination, respectively, as preferred initial treatments. The latter recommendation is based on recent trial evidence demonstrating the superior efficacy of a fixed-dose LABA/LAMA over an inhaled corticosteroid (ICS)/LABA in exacerbation prevention. ICS in combination with a LABA is also indicated for prevention of exacerbations, but the use of ICS is associated with an increased risk of adverse events such as pneumonia, and offers limited benefits beyond those provided by LABA or LAMA monotherapy. In this review, we examine evidence from a number of pivotal studies of LABAs and LAMAs, administered as monotherapy or as part of dual or triple combination therapy, with a specific focus on their effect on exacerbations. We also discuss a new proposed treatment paradigm for the management of COPD that takes into account this recent evidence and adopts a more cautious approach to the use of ICS. In alignment with GOLD 2017, we suggest that ICS should be reserved for patients with concomitant asthma or in whom exacerbations persist despite treatment with LABA/LAMA.


Background
The natural trajectory of chronic obstructive pulmonary disease (COPD) is punctuated by exacerbations, defined as an acute worsening of symptoms that results in additional therapy [1,2]. In many cases, exacerbations are triggered by respiratory tract infections (predominantly viral, but also bacterial) and environmental factors such as air pollution, yet in approximately one third of cases, the cause remains unknown [3].
COPD exacerbations have a marked negative effect on both the patient and underlying disease processes, and can result in hospitalization and readmission, an increased risk of death [4] and a significant reduction in health status [5,6]. Exacerbations are also associated with long-term decline in lung function and a high socioeconomic cost [7][8][9][10]. Thus, optimizing the prevention and management of COPD exacerbations is an important clinical issue.
A key step towards meeting this goal is to identify patients at greatest risk of exacerbation. The 'frequent exacerbator' phenotype (≥2 exacerbations/year) describes patients who are particularly susceptible to exacerbations, in contrast to infrequent exacerbators [11,12]. The exacerbator phenotype, which remains relatively stable over time [12], has a complex pathophysiology and is prevalent across all disease severities, but is more common in patients with worse lung function [12].
The Global Initiative for Chronic Obstructive Lung Disease (GOLD) provides treatment recommendations for patients who are at high risk of exacerbation: a longacting muscarinic antagonist (LAMA) or a long-acting β 2 -agonist (LABA)/LAMA combination are recommended as primary choice treatment for these patients. Although the GOLD strategy document was developed with an international audience in mind, some countries and regions (e.g. Spain, UK, South America, and Canada) follow guidance outlined in their own recommendations as well [13][14][15][16].
Inhaled corticosteroids (ICS) are recommended for patients at high risk of exacerbation with concomitant asthma, or who experience further exacerbations following initial bronchodilator treatment [2]. However, there is widespread evidence of inappropriate use of ICS in patients at low risk of exacerbation [17,18]. Considering the known risks associated with long-term ICS use, only patients carefully considered as suitable candidates should receive treatment with the appropriate ICS and associated dose, such that treatment benefit will outweigh any potential risk [19].
In this review, we will examine the efficacy of various COPD treatments in terms of exacerbation prevention, with particular focus on bronchodilators, and discuss a new proposed treatment paradigm for reducing exacerbation risk in patients with COPD.

Non-pharmacological intervention
Smoking cessation is the most effective initial strategy for reducing COPD disease progression, particularly exacerbation risk [2,35]. Smoking cessation significantly reduces the progressive decline in lung function [36], and substantially lowers the risk of mortality [37], however a significant number of COPD patients continue to smoke [38]. Although smoking cessation results in a reduced rate of decline in lung function, disease progression may persist [39].
Furthermore, a significant reduction in exacerbation rate has also been observed with vaccination versus placebo [40]. Findings from a population-based study suggested that COPD patients, especially the elderly, had a decreased risk of ischemic heart disease when receiving the influenza vaccine over many years [44].
Pulmonary rehabilitation (PR) reduces hyperinflation by promoting lung deflation and better lung ventilation, which is linked to an improved health status and exercise capacity [45]. PR also provides benefits that extend beyond the initial training period, such as: improved survival; improved recovery following hospitalization for exacerbation; reduced perceived intensity of breathlessness; and reduced anxiety and depression [2,46,47]. PR programs can significantly reduce the frequency of exacerbations and hospitalization, and the proportion of patients classified as 'frequent exacerbator' [48]. Non-pharmacological interventions, such as surgical modes or bronchoscopic modes of lung volume reduction, are also associated with a reduction of exacerbation; however, these interventions are limited to a small number of patients [49].
Encouraging patients to increase their levels of daily activity is also recommended, due to the general beneficial effects of physical exercise [2] and the reported links between lower levels of physical activity in patients with COPD and an increased risk of hospitalization [50,51] and readmission [52].

Pharmacological interventions
It must be recognized that many studies presented here were not designed to test for the relative efficacy of treatments in exacerbation prevention, making it difficult to draw conclusions on their effects in at-risk populations. However, the following studies were powered to test for differences in exacerbation rate, and recruited patients at high risk of exacerbation, thus, enriching study populations: POET (1 year) [53]; SPARK (64 weeks) [24]; FLAME (52 weeks) [29]; and INSPIRE (2 years) [54].

Single bronchodilation versus placebo in the prevention of exacerbations
The two main classes of bronchodilators are β 2 -agonists and muscarinic receptor antagonists. Both classes of bronchodilators improve patients' ability to breathe by relaxing airway smooth muscle, thereby reducing respiratory muscle activity and dynamic hyperinflation, and improving ventilatory mechanics [55][56][57]. Exacerbations are mainly triggered by infections associated with small airway inflammation [3], however environmental conditions, such as pollution, may also initiate or amplify these events [2]. The mechanisms by which bronchodilators prevent exacerbations are unclear, yet are thought to include decreased hyperinflation and mechanical stress, decreased mucus production and enhanced mucociliary clearance, the improvement of symptom severity and fluctuation, and potential anti-inflammatory properties [58]. Compared with placebo, the use of tiotropium 18 μg once daily (q.d.) was associated with sustained reductions in lung hyperinflation and inspiratory capacity both at rest and during exercise, contributing to improvements in exertional dyspnea and increased exercise endurance in patients with COPD [57]. Treatment with formoterol 12 μg twice daily (b.i.d.) significantly enhanced mucus clearance compared with tiotropium 18 μg q.d. in patients with mild-to-moderate COPD [59]. AUGMENT demonstrated that treatment with both formoterol 12 μg b.i.d. and aclidinium 400 μg b.i.d. significantly improved dyspnea and health status compared with placebo in patients with moderate-to-severe COPD [60]. Compared with formoterol 12 μg b.i.d., tiotropium 18 μg q.d. demonstrated a superior anti-inflammatory activity profile, significantly reducing the production of superoxide and pro-inflammatory mediators in COPD patients [61].
Traditionally, much of the evidence for the efficacy of LAMAs in exacerbation prevention has come from studies with tiotropium. More recently, there have been similar findings with glycopyrronium, aclidinium and umeclidinium, as well as the LABAs salmeterol and indacaterol. Table 1 summarizes the findings from major trials of these agents with regards to exacerbation prevention, although it should be noted that exacerbations were studied as a secondary endpoint in the majority of cases.
In two pivotal studies, GLOW1 (26 weeks) and GLOW2 (1 year), glycopyrronium (50 μg q.d.) significantly prolonged time to first moderate-to-severe exacerbation versus placebo [71,72]. In GLOW1, glycopyrronium also significantly reduced the risk of severe COPD exacerbations leading to hospitalization and the proportion of hospitalizations due to COPD exacerbations [71]. In GLOW2, glycopyrronium significantly reduced the rate of moderate-to-severe exacerbations and the number of exacerbations requiring treatment with systemic corticosteroids or antibiotics, versus placebo [72].
In ACCORD (12 weeks) and ATTAIN (24 weeks), aclidinium (200 or 400 μg b.i.d.) significantly reduced the rate of exacerbations of any severity and numerically reduced rates of moderate or severe exacerbations per patient per year compared with placebo [73,74]. Two 24-week studies examining the efficacy of umeclidinium demonstrated significant reductions in the risk of exacerbations versus placebo [22,75].

Comparison of the efficacy of single bronchodilators in the prevention of exacerbations
Only a few head-to-head studies have examined the relative effects of different bronchodilators on exacerbation outcomes ( Table 2).
The first study to specifically test the efficacy of a LABA versus a LAMA in exacerbation prevention was POET, a randomized, double-blind, double-dummy, parallel-group trial in patients with moderate-to-very-severe COPD and a history of exacerbations. Compared with salmeterol ( [80]. To date, no direct head-to-head, LAMA versus LAMA studies have been performed. In GLOW2, open-label tiotropium (18 μg q.d.) was included as a reference arm; however, the study was not designed nor powered to test for differences between the two active treatments. Compared with placebo, exacerbation risk was reduced with glycopyrronium (HR 0.66; p = 0.001) and with tiotropium (HR 0.61; p = 0.001), although no formal comparisons between the two treatments were made [72]. In SPARK, which studied the efficacy of indacaterol/glycopyrronium (IND/GLY 110/50 μg q.d.) compared with glycopyrronium (50 μg q.d.) and tiotropium (18 μg q.d.) in patients with severe-to-very severe COPD and an exacerbation history, the rate of moderate or severe exacerbations was similar between glycopyrronium and tiotropium monotherapies (HR 1.03; p = 0.68) [24].

Dual bronchodilation versus placebo in the prevention of exacerbations
The mechanisms underlying interactions between LABAs and LAMAs have not been fully elucidated. However, β 2agonists can amplify muscarinic antagonist-mediated smooth muscle relaxation by modulating cholinergic neurotransmission and decreasing acetylcholine release, and muscarinic antagonists can augment β 2 -agonist-mediated bronchodilation by reducing the bronchoconstrictor effects of acetylcholine [81]. The complimentary mechanisms of action of LABAs and LAMAs elicit additive effects on lung function, and provide a rationale for combining the two agents to optimize bronchodilation. Mechanisms that most likely involve reduced airway resistance, improved inspiratory capacity and reduced hyperinflation may confer benefits in terms of exacerbations [58].
Numerous dual bronchodilators have been developed with this aim in mind. They include once-daily IND/ GLY, umeclidinium/vilanterol (UMEC/VI) and tiotropium/olodaterol (TIO/OLO), and twice-daily aclidinium/formoterol (A/F) and glycopyrrolate/formoterol fumarate (GLY/F). In the USA, IND/GLY has been developed for twice-daily use [82]. Table 1 summarizes data from studies comparing the effects of dual bronchodilators on exacerbations with placebo.
Exacerbation risk was significantly reduced relative to placebo in UMEC/VI (62.5/25 and 125/25 μg) studies [22,75] whereas the effects of A/F were less consistent across different doses (400/12 and 400/6 μg) and exacerbation assessments [23,83]. In a prespecified analysis of pooled data from ACLIFORM and AUGMENT, the higher dose of A/F significantly reduced the rate of moderate or severe exacerbations compared with placebo, whether defined according to EXAcerbation of Chronic Pulmonary Disease Tool (EXACT) criteria (RR 0.78; p < 0.01) [84] or healthcare resource utilization (HRU; RR 0.71; p < 0.05) [83]. The higher-dose combination of A/F also prolonged time to first exacerbation of any severity defined according to HRU (HR 0.72; p < 0.05) or EXACT (HR 0.79; p < 0.05) versus placebo ( [83]. There are currently no published data comparing IND/GLY or TIO/OLO with placebo [23,83].

Dual versus single bronchodilation in the prevention of exacerbations
As shown in Table 2, LABA/LAMAs can improve exacerbation outcomes compared with monotherapy, although not all studies were designed for this objective, and the results are variable. In SPARK, an exacerbation study, the annualized rate of moderate or severe exacerbations was significantly lower with IND/GLY versus glycopyrronium (primary endpoint, RR 0.88; p = 0.038), and rates of all exacerbations (mild, moderate or severe) were significantly lower with IND/GLY versus either glycopyrronium (RR 0.85; p = 0.0012) or tiotropium (RR 0.86; p = 0.0017) [24].

Single bronchodilation versus ICS/LABA combinations in the prevention of exacerbations
Of the six ICS/LABA versus bronchodilator monotherapy studies, only two demonstrated significant benefits on exacerbations (Table 3) [87,88]. It should be noted that these studies were not all designed to compare the effects of an ICS/LABA with single bronchodilation (but with placebo), and exacerbation endpoints were often secondary or exploratory.
One of the most robust studies was TORCH (3 years), which studied deaths (any cause) as the primary outcome, as well as exacerbation frequency between treatments. In TORCH, salmeterol/fluticasone propionate combination (SFC 50/500 μg b.i.d.) significantly reduced the annual rate of moderate or severe exacerbations compared with salmeterol (50 μg b.i.d.; RR 0.88; p = 0.002) [87].

Dual bronchodilation versus ICS/LABA combinations in the prevention of exacerbations
Exacerbation data are available from a number of studies comparing a LABA/LAMA with an ICS/LABA (Table 3).    pooled data from LANTERN and ILLUMINATE, the annualized rate of moderate or severe exacerbations was significantly lower with IND/GLY versus SFC, in both the whole population (p = 0.02) and in subgroups of patients classified as either GOLD Group B (p = 0.16) or GOLD Group D (p = 0.05; patients were classified according to GOLD 2009 and 2010 for ILLU-MINATE and LANTERN, respectively). Furthermore, IND/GLY delayed the time to first moderate or severe exacerbation compared with SFC in the overall population and in GOLD Group B and GOLD Group D subgroups [91].
The most recent study comparing a LABA/LAMA with an ICS/LABA was FLAME, which specifically studied the differences in exacerbations between IND/GLY (110/50 μg q.d) and SFC (50/500 μg b.i.d.) as the primary outcome, and included an enriched patient population at high risk of exacerbation (≥1 exacerbation in the previous year) [29]. Compared with SFC, IND/GLY significantly reduced the rate of all COPD exacerbations (RR 0.89; p = 0.003), and the rate of moderate or severe exacerbations (RR 0.83; p < 0.001). Additionally, IND/ GLY significantly prolonged the time to first exacerbation (HR 0.84; p < 0.001). The time to first moderate or severe exacerbation (HR 0.78; p < 0.001), and time to the first severe exacerbation (HR 0.81; p = 0.046) were also significantly prolonged with IND/GLY versus SFC. Treatment benefit or time to first exacerbation was detected as early as 4 weeks. Compared with SFC, IND/ GLY numerically, but not significantly, reduced the rate of exacerbations in patients with a history of ≥2 exacerbations in the previous year, (19% of the patient population). However, it is worth noting that this was a subgroup analysis and FLAME was not powered to detect treatment differences in subgroups [29].
In a 24-week trial comparing A/F (400/12 μg b.i.d.) with SFC (50/500 μg b.i.d), a similar proportion of patients experienced at least one exacerbation in the A/F and SFC groups, regardless of exacerbation definition (HRU or EXACT criteria) [27]. However, the study was not designed to test exacerbations.

Triple therapy in the prevention of exacerbations
Evidence for the efficacy of triple therapy (ICS/LABA/ LAMA) in exacerbation prevention is currently limited. Nevertheless, in TRILOGY (1 year), the adjusted annual rate of moderate-to-severe exacerbations was significantly reduced following step-up to triple therapy (beclomethasone dipropionate/formoterol fumarate/glycopyrronium bromide [ [94]. Findings are eagerly anticipated from the ongoing InforMing the PAthway of COPD Treatment (IMPACT) study, comparing the rate of moderate and severe exacerbations between FF/UMEC/VI and FF/VI or UMEC/VI over 52-weeks in symptomatic COPD patients with an exacerbation in the previous 12 months [95].

Mucolytics in the prevention of exacerbations
Mucolytics are oral medicines designed to reduce mucus and sputum viscosity, thereby making it easier for patients to cough up mucus and clear it from the airways [96]. The current GOLD recommendations recognize that regular use of mucolytics may be beneficial in patients not receiving ICS, to reduce exacerbations and improve health status [2].
The antioxidant N-acetylcysteine (NAC; 600 mg/day), is recommended by the American College of Chest Physicians and Canadian Thoracic Society for patients with moderate-to-severe COPD and a history of two or more exacerbations in the previous 2 years [97]. However, NAC 600 mg/day is currently not approved for use in the USA. In a 6-month study, the number of exacerbations was reduced by 41% with standard therapy plus NAC versus standard therapy alone, and fewer patients experienced at least one exacerbation (46 versus 63 patients, respectively) [98]. In PANTHEON (1 year), NAC (600 mg/b.i.d.) significantly reduced the annual rate of exacerbations versus placebo (1.16 versus 1.49 exacerbations per patient-year; p = 0.0011) [99]. By contrast, in BRONCUS (3 year), a randomized, placebocontrolled study, there was no difference between NAC (600 mg/day) and placebo in the number of exacerbations per year (a primary outcome, [1.25 versus 1.29; p = 0.85]), although sub-group analyses suggested that NAC might have reduced exacerbation rate in patients not receiving ICS treatment [100]. In a 2015 meta-analysis of 13 studies (N = 4155), NAC significantly reduced the relative risk of exacerbations in patients with COPD and/or chronic bronchitis. The authors concluded that a dose of 600 mg b.i.d. should be used in patients who have both chronic bronchitis and COPD, whereas the standard 600 mg/day dose should suffice for patients who have chronic bronchitis alone [101]. Accordingly, one study showed that NAC (600 mg b.i.d.) was more effective than placebo in reducing exacerbation risk and prolonging time to first exacerbation in high-risk patients (GOLD Groups C and D, according to the 2011 GOLD recommendations), but not low-risk patients [102].
Two studies have examined the effects of erdosteine on exacerbations. EQUALIFE, an 8-month, randomized, double-blind trial, demonstrated that patients taking erdosteine (300 mg b.i.d.) had significantly fewer exacerbations and spent fewer days in hospital than those on placebo [103]. More recently, Moretti et al demonstrated that 10-day treatment with erdosteine (900 mg/day) was associated with a 39% lower risk of exacerbations in the 2 months post-discharge and a significant delay in time to first exacerbation at post-discharge days 30 (p = 0.009) and 60 (p = 0.075) compared with placebo, in patients hospitalized following acute exacerbation [104]. In a 6-month, randomized trial, a significantly higher proportion of patients experienced no exacerbations when treated with continuous carbocysteine lysine salt monohydrate (SCMC-Lys; 2.7 g q.d.) compared with placebo (p < 0.001) [105]. Similarly, in PEACE (1 year), the annualized rate of exacerbations was significantly lower with carbocysteine (1500 mg/day) than with placebo, representing a 25% reduction in risk (p = 0.004) [106].

Phosphodiesterase-4 inhibitors in the prevention of exacerbations
Phosphodiesterase-4 (PDE-4) inhibitors can inactivate immune and inflammatory cells by blocking the metabolism of cyclic adenosine monophosphate (cAMP) [107]. GOLD recommends addition of the selective, long-acting PDE-4 inhibitor roflumilast to a ICS/LABA/LAMA in GOLD Group D patients who continue to experience exacerbations despite triple therapy, particularly patients with a forced expiratory volume in one second (FEV 1 ) <50% predicted, chronic bronchitis, and ≥1 hospitalization for an exacerbation in the previous year [2].
A meta-analysis of 13 studies suggested that roflumilast (500 μg q.d.) was more effective than placebo in reducing the rate of acute exacerbations (p < 0.001) [107]. In REACT, roflumilast (500 μg q.d.) reduced the rate of moderate-to-severe exacerbations by 13.2% versus placebo in patients with severe COPD, chronic bronchitis and at risk of frequent and severe exacerbations, and receiving ICS/LABA treatment with or without tiotropium (Poisson regression analysis, p = 0.0529) [108]. In a post-hoc analysis of RE(2)SPOND (52 weeks), roflumilast (500 μg q.d.) significantly reduced the rate of moderate or severe exacerbations versus placebo in patients with severe-to-very severe COPD and chronic bronchitis, a history of >3 exacerbations and/or ≥1 hospitalizations in the prior year [109]. Roflumilast is indicated as a maintenance treatment (added on to bronchodilator therapy) in adults with severe COPD associated with chronic bronchitis, and a history of frequent exacerbations [110].

Macrolides in the prevention of exacerbations
Macrolides are antibiotics with antimicrobial, antiinflammatory, and immunomodulating effects. GOLD 2017 recommends the addition of a macrolide to an ICS/LABA/LAMA regimen in GOLD Group D patients who are former smokers and continue to suffer exacerbations despite triple therapy [2]. In COPD, the best studied macrolide is azithromycin. COLUMBUS (1 year) demonstrated that exacerbation rate was significantly reduced with azithromycin (500 mg three times/week), compared with placebo (odds ratio = 0.58, p = 0.001), in patients with COPD who had received treatment for ≥3 exacerbations in the previous year despite optimal inhalation therapy [111]. Similar findings were reported in a 1-year, randomized, controlled trial in patients with COPD at risk for exacerbations, where azithromycin (250 mg q.d.) significantly delayed median time to first exacerbation (266 versus 174 days; p < 0.001), and significantly reduced the frequency of exacerbations (1.48 versus 1.83; p = 0.01) versus placebo [112]. Similar results have been reported in other, smaller studies comparing erythromycin and placebo [113,114].

Appropriate use of ICS
Much evidence supports the use of ICS in patients with persistent asthma, yet the role of ICS in preventing exacerbations of COPD is less clear [19,115]. Various methodological issues in trial design and/or statistical analysis affect results and make subsequent study interpretation difficult [116].
However, data are emerging to suggest that raised blood or sputum eosinophil levels could predict a positive response (i.e. a reduction in exacerbations) to ICS/ LABA versus LABA monotherapy, or predict any deleterious effects ICS withdrawal may have [117][118][119][120][121].
The potential association between this 'eosinophilic phenotype' and a positive response to ICS requires further investigation, as the effect has not been consistent across studies within the same analysis [118], and may only be present in patients with a history of ≥2 exacerbations in addition to raised eosinophil levels [121]. Prospective analysis of FLAME (which excluded patients with a previous diagnosis of asthma and/or a blood eosinophil count >600/mm 3 ), demonstrated that IND/GLY was superior to SFC in reducing the rate of moderateto-severe exacerbations, regardless of baseline eosinophil levels [29]. Thus, further studies are needed before any recommendations can be made regarding the potential use of ICS in specific sub-populations.
Proposed Treatment Paradigm Figure 1 shows a proposed treatment paradigm for exacerbation prevention based on the evidence presented and centered on optimal use of bronchodilation, which has previously been published [122][123][124].

Conclusions
Not only do COPD exacerbations negatively impact the underlying disease course, but they also have a detrimental effect on patients' lives, resulting in lung function decline, increased risk of mortality and poor health status. While there has been a tendency to recommend the use of ICS for patients at high risk of exacerbations, ICS are associated with a myriad of side effects such as pneumonia. However, (and as recognized in the updated GOLD strategy document) evidence is emerging that suggests there may be more appropriate treatment strategies for many at risk patients, including a LABA/ LAMA combination.
The proposed treatment paradigm for exacerbation prevention is centered on optimizing bronchodilation as an initial pharmacological step, first with a LAMA, and subsequently with a dual LABA/LAMA should symptoms Fig. 1 A proposed treatment algorithm for the treatment of chronic obstructive pulmonary disease (COPD). FEV 1 , forced expiratory volume in 1 s; GOLD, Global Initiative for Chronic Obstructive Lung Disease; ICS, inhaled corticosteroids; LABA, long-acting β 2 -agonist; LAMA, long-acting muscarinic antagonist; PDE-4, phosphodiesterase-4 [134] worsen or exacerbation occur. Only if patients continue to suffer exacerbations do we suggest the addition of an ICS or a PDE-4 inhibitor, depending on patient profile/phenotype. It is possible that a subgroup of patients with COPD who have raised blood or sputum eosinophils may respond better than others to ICS, although current data are still preliminary and somewhat contradictory. Future studies are warranted to better define the groups who may benefit from ICS, and to identify the mechanisms by which bronchodilation reduces exacerbations.