Inhaled corticosteroids in COPD: Benefits and risks

Introduction

The Global Burden of Disease Study has recently reported that Chronic Obstructive Pulmonary Disease (COPD) is the 3rd leading cause of death worldwide, reaching this threshold about ten years earlier than predicted.^1,2 Explanations for this rise are multifactorial and include the aging population and improved outcomes for cardiovascular (CV) disease and cancer. Additionally, no pharmacologic treatment has been definitively shown to improve COPD mortality which also contributes. Other treatment goals for COPD include reducing the burden of disease by alleviating symptoms, improving health-related quality of life, reducing exacerbation frequency and severity, and preservation of lung function.^3,4 Achieving these goals is challenging due to the heterogeneity and complex pathobiology of the disease which includes dysregulation in immune activation, airflow obstruction due to bronchoconstriction, small airway fibrosis, emphysema and muco-obstruction, as well as dysfunction of the mucociliary escalator.^5,6 The presence of lung inflammation also promotes airway and alveolar tissue damage, airway remodeling, and airflow obstruction and contributes to the risk of exacerbations of COPD (ECOPD) which cause life-limiting symptoms and are also associated with loss of lung function, poor quality of life and significant healthcare costs.^2,7–9

Anti-inflammatory therapeutics evaluated in COPD include inhaled corticosteroids (ICS), oral glucocorticoids, phosphodiesterase inhibitors, antibiotics, statins, mucolytics, and monoclonal antibodies targeting inflammatory mediators such as benralizumab and mepolizumab that target IL-5, among other drugs.^5,10,11 The best studied of these approaches is the use of ICS, alone and in combination with a long-acting beta-agonist (LABA), and with long-acting muscarinic antagonists (LAMA). Though targeting lung inflammation with ICS can have clinical benefits, including reducing the risk of exacerbations, it can also be associated with untoward effects.

ICS Monotherapy

Though ICS therapy has a well-established role in the management of asthma, early studies in COPD patients were small and resulted in conflicting results.^12–15 Subsequent more extensive studies suggested modest benefits on lung function, symptoms, and exacerbation risk but did not demonstrate effects on the rate of FEV1 decline and raised the possibility of a mortality disadvantage with long term use.^16–18

Paggiaro and colleagues reported that six months of monotherapy with fluticasone propionate resulted in increased 6-minute walking distance, improved symptom scores for cough and sputum production, as well as a reduction in the number of patients who suffered moderate or severe exacerbations (60% vs. 86%, P<0.001) compared with placebo.¹⁶ Pauwels et al. reported that in COPD patients who continued to smoke, monotherapy with budesonide resulted in initial improvement in FEV₁ in the first six months of treatment compared with placebo, but that the rate of decline in FEV₁ over three years in the two groups was similar.¹⁷ These studies were followed by a three-year single-center, double-blind, randomized placebo-controlled study nested in the Copenhagen City Heart Study. In this trial, Vestbo et al. evaluated the effect of budesonide on the rate of FEV₁ decline in COPD patients and again found no clear treatment benefit at three years with ICS monotherapy.¹⁹ The Inhaled Steroids in Obstructive Lung Disease in Europe (ISOLDE) study also reported no statistical difference in decline in FEV₁ with fluticasone propionate compared with placebo, though there was a 25% reduction in median annual exacerbation rate and lower decline in health status as measured by St George’s respiratory questionnaire (SGRQ) with ICS therapy.²⁰ The Towards a Revolution in COPD Health (TORCH) trial randomized over 6,000 patients with COPD to fluticasone propionate/salmeterol (FP/SAL), salmeterol (SAL), fluticasone propionate (FP), or placebo. FP monotherapy was again shown to reduce the rate of exacerbations versus placebo, but the risk of death at three years was actually increased [15.2% vs. 16.0% (HR 1.060, 95% CI 0.886–1.268, P=0.53)], and statistically significantly higher when compared with FP/SAL years [16.0% vs. 12.6%; P=0.007]. Though there have been additional studies of ICS monotherapy in subpopulations of COPD, including those with heightened cardiovascular risk, and the suggestion that they may impact lung function decline in those with elevated blood eosinophils, the results of the TORCH study strongly suggested that this strategy should not be used routinely.^21,22

ICS/LABA Therapy

Numerous studies have compared ICS/LABA to LABA or LAMA alone as well as to dual bronchodilator therapy and have examined a range of clinical outcomes, though the effects of ICS/LABA on exacerbations and mortality have been of greatest interest.

Exacerbations:

ICS/LABA vs. LABA:

In TORCH, FP/SAL was associated with a reduction in moderate or severe ECOPD when compared to placebo, FP, and SAL, the latter being the most clinically relevant comparison with a risk ratio of 0.88 (95% confidence [CI], 0.81–0.95) favoring combination treatment. Though there are differences in the ICS tested, as well as trial methodology and the populations enrolled, ICS/LABA has been repeatedly shown to reduce the rate of ECOPD by 10–25%.^18,23–36

ICS/LABA vs. LAMA:

The Investigating New Standards for Prophylaxis in Reducing Exacerbations (INSPIRE) trial evaluated tiotropium therapy compared to FP/SAL for patients with severe and very severe COPD (mean FEV₁ 39%) over two years and found no difference in annual exacerbation rate (rate ratio [RR] 0.967; 95% CI, 0.836–1.119; P = 0.656). FP/SAL was associated with better health status and a lower withdrawal rate, suggesting an overall advantage to ICS/LABA.

ICS/LABA vs. LAMA/LABA:

The Indacaterol–Glycopyrronium versus Salmeterol–Fluticasone for COPD (FLAME) trial demonstrated the superiority of the LABA/LAMA for the endpoint of annual rate of moderate or severe COPD exacerbations (RR 0.83; 95% CI, 0.75–0.91; P<0.001). These results contrast with those from the IMPACT study which showed an exacerbation benefit of fluticasone furoate/vilanterol (FF/VI) compared with umeclidinium/vilanterol (UMEC/VI). This difference in outcome is likely explained by methodological differences between the trials, including the inclusion of a run-in period on tiotropium in FLAME, and the fact that the population in IMPACT was at much higher exacerbation risk with 54% have ≥ two moderate to severe exacerbations and 26% with ≥ one severe exacerbation requiring hospitalization.^37,38

Survival/Mortality:

In TORCH, FP/SAL was associated with a trend towards a reduction in mortality (HR 0.825, 95% CI 0.68–1.00, P=0.052). Though the results did not reach statistical significance, a prespecified secondary analysis using cox proportional hazard testing did suggest a survival advantage with FP/SAL compared with placebo (HR 0.811 (95% CI, 0.670–0.982; P=0.03).¹⁸ There are several possible explanations for the study failing to definitively establish a mortality benefit including lower than expected mortality and differential dropout in the placebo arm.³⁹

A post hoc analysis of the TORCH study did demonstrate that in those patients treated for cardiovascular disease and with moderate airflow limitation (FEV₁ > 50%), FP/SAL reduced the risk of cardiovascular adverse events compared with placebo. These data suggested the possibility of ICS/LABA being especially beneficial to those with heightened cardiovascular risk.⁴⁰ This hypothesis was tested in the Study to Understand Mortality and Morbidity in COPD (SUMMIT) study, that enrolled patients with moderate COPD and heightened cardiovascular risk and randomized them to fluticasone furoate (FF), vilanterol (VI), their combination (FF/VI), or placebo. The study was designed to evaluate the effect of FF/VI on all-cause mortality, but cardiovascular events and exacerbations were also examined. All-cause mortality was unaffected by FF/VI compared with placebo (HR 0.88, 95% CI 0·74–1·04) and though there was also no impact of any active treatment on CV outcomes, there was no evidence of an adverse effect.⁴¹ Prespecified secondary analyses demonstrated that FF/VI reduced the risk of exacerbations compared with placebo and that FF containing treatments might slow the rate of FEV1 decline.^42,43 The study also demonstrated an important link between exacerbations and cardiovascular events as the hazard ratio for cardiovascular events after ECOPD was increased, particularly in the first 30 days (HR 3.8; 95% CI, 2.7–5.5) and remained elevated for up to 1 year.⁴⁴

The INSPIRE trial examined survival as a safety endpoint and demonstrated that FP/SAL was associated with a reduction in mortality compared with tiotropium (3% versus 6%; P=0.032). Additionally, cox proportional hazards model for time to death on treatment showed a 52% reduction in the risk of death in those randomized to FP/SAL (HR 0.48, 95% CI, 0.27–0.85; P = 0.012)^45,46

Trials of Triple Therapy with ICS/LABA/LAMA

The recent development of single-inhaler triple therapies has prompted a series of clinical trials testing the effects of maximal inhaled treatment with an emphasis on exacerbations.

Exacerbations

Triple therapy vs. LAMA/LABA:

The InforMing the Pathway of COPD Treatment trial (IMPACT) evaluated the effect of triple therapy with FF/VI/umeclidinium (UMEC) versus both FF/VI and UMEC/VI in more than 10,000 patients with symptomatic COPD and a history of exacerbations.³⁸ There was a lower rate of moderate and severe COPD exacerbations with triple therapy compared with UMEC/VI (RR 0.75; 95% CI, 0.70–0.81; P<0.001), as well as a lower rate of COPD-related hospitalizations (RR 0.66; 95% CI, 0.56–0.78; P<0.001). The efficacy and safety of fixed triple inhaler therapy was also demonstrated in TRILOGY as well as TRINITY where the combination of Beclometasone Dipropionate, Formoterol Fumarate and Glycopyrronium (BDP/FF/G) showed a 23% reduction in moderate/severe exacerbations vs. ICS/LABA (BDP/FF) (RR 0.77, 95% CI 0.65–0·92; P=0.005) and 20% reduction vs LAMA (Tiotropium) (RR 0.80, 95% CI 0.69–0.92: P=0·0025), in symptomatic severe/very severe COPD patients with an exacerbation history.^47,48

Similar to IMPACT, the Extrafine Inhaled Triple Therapy versus dual Bronchodilator Therapy in Chronic Obstructive Pulmonary Disease (TRIBUTE) study also compared triple ICS-containing therapy versus LABA/LAMA in patients with severe/very severe COPD at risk for ECOPD.^38,49 BDP/FF/G reduced the risk of moderate/severe exacerbations compared with indacaterol/glycopyrronium (IND/GLY) (annual exacerbation rate of 0.50 versus 0.59 events per patient-year; RR 0.848: 95% CI 0.723–0.995, P=0.043). Though the overall rate of exacerbations and the relative benefit of triple therapy was less than observed in IMPACT, the results still suggest a clinically relevant advantage.⁴⁹

Survival/Mortality:

In IMPACT, triple inhaled therapy with ICFF/VI/UMEC was associated with a signal towards a reduction in all-cause mortality compared with UMEC/VI. While assessment of the mortality signal in IMPACT should be interpreted with caution since it was not the primary outcome, the trial is invaluable because of its meticulous adjudication of clinical events and the assessment of outcomes both on and off treatment and after capture of the vital status on all but 42 (0.4%) patients. With time to on-treatment all-cause mortality as a prespecified endpoint, the hazard ratio for FF/VI/UMEC versus UMEC/VI was 0.58 (95% CI, 0.38–0.88; unadjusted P = 0.011). Similar results were observed when the analysis included data after patients stopped treatment (HR 0.71; 95%CI, 0.51–0.99; P=0.043) and when the vital status was available for 99.6% of randomized subjects (HR 0.72; 95%CI, 0.53–0.99, P=0.042).^38,46 A pooled analysis of the TRILOGY, TRINITY and TRIBUTE studies evaluated mortality as a safety outcome and found that the risk of non-respiratory mortality was lower with ICS-containing treatments compared to treatment without ICS (HR 0.65, 95%, CI 0.43–0.97; P=0.037)^47–51

Though no study has demonstrated an advantage to ICS-containing therapy with mortality as the primary endpoint, the weight of the evidence suggests that there may be a benefit in a subset of patients, likely those with more severe disease, significant symptoms, and exacerbation risk.

Risks/Safety

Population-based studies and analyses of RCT have shown that ICS therapy has untoward effects that include changes in bone density, oral candidiasis, insulin resistance, skin changes, adrenal suppression, as well as increased risk of pneumonia and mycobacterial infections (Table 1).^52–59

Table 1.

Side-effects of inhaled corticosteroids in COPD and strength of evidence

	Randomized Controlled Trial	Observational/Case- Control Study	Systematic Review
Pneumonia	✓✓✓	✓✓✓	✓✓✓
Mycobacterial Infections		✓	✓
Diabetes		✓
Bone density	(No effect on Fracture Risk)	✓✓	✓✓

Pneumonia Risk:

COPD patients are at increased risk of community-acquired pneumonia as well as increased risk of death from pneumonia.⁶⁰ Risk factors for pneumonia in COPD include older age, lower BMI, male gender, lower FEV1, prior pneumonia, and ICS use, though the magnitude of additional risk attributable to ICS is uncertain.⁶¹ There has also been substantial interest in the use of blood eosinophils as a predictor of pneumonia, but most studies have shown either no association or an association only with very low eosinophil counts.^62,63 There is also significant variability in pneumonia rates across studies with differences in study populations, study design, the ICS included, and varied definitions of pneumonia that complicate comparisons. Despite these challenges, a Cochrane review has concluded that pneumonia is a class-related risk of ICS.⁶⁴

The TORCH study reported increased pneumonia risk in the FF containing treatment groups, and while this was not a pre-specified outcome and chest radiographs were not required for diagnosis, the same trend has been observed in subsequent RCT’s of ICS containing therapies, including those where pneumonia events were adjudicated.^18,38,65

It is important to note that while the weight of the evidence suggests that ICS increase the risk of any pneumonia, this does not seem to increase pneumonia-related or all-cause mortality. There is also data to suggest that outcomes for COPD patients admitted with pneumonia are better in those taking ICS as outpatients and that there may be a double effect of the drugs including both an adverse effect and an additional unidentified mitigating effect.⁶⁶ Additionally, in order to determine overall risk-benefit, the risk of ICS-related pneumonia has to be compared and counterbalanced with the ICS-related reduction in acute exacerbations. Data from IMPACT³⁸ and TRIBUTE⁴⁹ show that the rate of ECOPD is approximately ten-fold higher than the rate of pneumonia and thus, the small increase in pneumonia risk with ICS may be offset by the exacerbation benefit though this must the decision to use ICS must be individualized.

Mycobacterial Risk:

Recent population-based studies suggest that COPD patients on ICS are at an increased risk of non-tuberculous mycobacterial (NTM) infections.^67,68 In a large population-based study in Canada, Brode and colleagues found that there was an association between NTM and ICS use in patients with obstructive lung disease (asthma, COPD, asthma-COPD overlap). When selected for COPD patients, the authors found an adjusted OR of 1.96 (95% CI, 1.62–2.36).⁶⁸ Similar findings were seen in a population-based case-control study in Denmark that verified that COPD is a risk factor for NTM, and ICS therapy further increases the risk.⁶⁷ This effect of ICS on mycobacterial risk has not been demonstrated in a randomized clinical trial.

Fractures and Bone Density:

A longstanding concern for ICS therapy is an effect on bone density and fracture risk. Osteoporosis is significant comorbidity in COPD, and patients not on ICS therapy are estimated to have a prevalence of ranging from 10% to 33%.⁶⁹ The systemic bioavailability of most ICS molecules is believed to be minimal, but the effects of very long exposure have not been adequately studied. In the Lung Health Study, after three years of ICS therapy with triamcinolone, a significant number of patients had reduced bone density in the femoral neck and lumbar spine compared to placebo.⁷⁰ These findings were not observed in other studies, and no RCT has demonstrated a significant increase in fracture risk.^71–74,73 The effects of ICS on bone density are likely both dose and time-dependent and many patients started on the drugs will take them for far longer than has been studied. It is therefore advisable that ICS dose and duration be minimized, therapy limited to those that will receive benefit, and individuals monitored for adverse bone effects.

Glycemic Control/Insulin Resistance:

Diabetes is a common comorbid condition in patients with COPD and thus understanding the effect of ICS on glycemic control is essential. A case-control study in patients with asthma or COPD reported that ICS use was associated with an increased risk of new-onset diabetes and diabetes progression, with a clear dose-response relationship.⁷⁵ In a UK based historical matched cohort study, Price et al. evaluated the change in HbA_1c in COPD patients with co-existing diabetes who were taking ICS over 12–18 months relative to a non-ICS cohort. The adjusted difference in change in HbA_1c was 0.16% (95% CI; 0.05–0.27%) suggesting that patients using ICS had worse diabetes control.⁷⁶ Another cohort study noted small changes in serum glucose (ranging from 2mg/dL to 5mg/dL) without an increased risk of developing diabetes even though the clinical relevance of this effect is unclear.⁷⁷ Additional studies are needed to determine the overall impact of ICS on long term glycemic control.

Adrenal Suppression:

Systemic administration of glucocorticoids causes hypothalamic-pituitary-adrenal (HPA) axis suppression by reducing corticotropin (ACTH) production, which reduces cortisol secretion by the adrenal gland. The degree of HPA suppression is dependent on dose, duration, frequency, and timing of glucocorticoid administration, however, there has been no data to suggest that doses delivered with inhaled corticosteroids have a meaningful clinical impact on the hypothalamic-pituitary-adrenal axis for most patients.^78–80

Risk-Benefit Considerations

The selection of patients for ICS-containing therapy requires an assessment of their likelihood of receiving a benefit, most clearly in reducing exacerbation risk and the risk of hospitalization, and their propensity to suffer ICS-related side effects. As discussed above, older age, low BMI, lower lung function, and male sex increase the risk of pneumonia in COPD but may not preclude treatment with ICS if the potential for exacerbation prevention is more significant. Several studies suggest that sputum eosinophilia, while not readily available in most clinical settings, may serve as a predictive biomarker of response to corticosteroids in patients with COPD.^4,81–84 While sputum eosinophils correlate well with exacerbations and ICS benefit, findings from the Subpopulations and Intermediate Outcome Measures in COPD Study (SPIROMICS) study suggest that the use of blood eosinophils as a surrogate may perform poorly due to weak association between sputum and blood eosinophilia.⁸⁵ This contrasts with data from several RCTs that demonstrate a clear association between greater eosinophil count and both the risk of exacerbations and the magnitude of ICS benefit. A post-hoc analysis of pooled phase III data also showed that patients with a peripheral blood eosinophil count of at least 100 cells/μL with at least one exacerbation in the preceding year had a reduction in exacerbations with ICS/LABA therapy (rate ratio 0.75, 95% CI 0.57–0.99, P=0.015) with the most significant treatment effect noted in those with the highest level of blood eosinophils.^85,86 IMPACT and TRIBUTE had a pre-defined analysis based on blood eosinophilia that showed more significant treatment effects to ICS with eosinophil level >150–200 cells/μL, though in smokers the threshold for benefit may be higher.^38,49 Pascoe and colleagues re-demonstrated a direct relationship between ECOPD reduction and higher peripheral eosinophil count however there was not a clear association between pneumonia with baseline blood eosinophils.⁸⁷ Based on available data, it is reasonable to consider ICS use in those patients with frequent and severe exacerbations and consider avoiding in patients with a history of frequent cases of pneumonia. There may also be a role for examining sputum culture data to select ICS candidates, as chronic bronchial infection may also increase pneumonia risk.⁶³

ICS withdrawal

There have been concerns about the risk of abrupt discontinuation of ICS in patients who are tolerating the drugs. During the run-in period of the ISOLDE trial, there was an increased rate of ECOPD following the withdrawal of ICS.²⁰ The Withdrawal of Inhaled Steroids during Optimized Bronchodilator Management (WISDOM) study assessed response to de-escalation of ICS therapy in patients on triple inhaled therapy. When the ICS dose was gradually decreased over 12 weeks, there was not an increased risk of exacerbations compared with the group that remained on triple ICS-containing therapy (HR 1.06, 95% CI 0.94–1.19). There was a small decline in FEV₁ (approximately 40mL) and mild worsening in health status in the ICS withdrawal group.⁸⁸ Subgroup analysis showed that the subset of the study population that had eosinophil counts > 4% had an increased risk of exacerbations when ICS were withdrawn (HR 1·63; 95% CI 1·19–2·24), and a similar pattern was noted for an absolute eosinophil count > 300 cells/μL.^88–93 The recent SUNSET trial confirmed that the patients that fair best with ICS withdrawal are those with stable disease, low eosinophils, and infrequent exacerbations.⁹⁴

Discussion

As more is understood about COPD and its varying endotypes, it is becoming increasingly clear that a standardized approach to this heterogeneous disease is inadequate and that it is essential to personalize care. ICS-based therapy continues to play an important role in COPD care as it results in improvements in lung function, healthcare-related quality of life, and reduced exacerbations compared with bronchodilator treatment alone. Triple therapy with ICS/LAMA/LABA may offer further benefit including a potential survival advantage in select patients. It is evident however that these ICS containing regimens are associated with adverse effects with pneumonia garnering the most attention. Factors associated with highest pneumonia risk are male sex, low BMI, low lung function, and advanced age but should not preclude ICS treatment if the exacerbation benefits exceed this risk. Practical factors to assist in determining patient suitability for long term pharmacotherapy with ICS containing compounds should include evaluating exacerbation history, including the frequency and severity of prior events, blood eosinophils, history of pneumonia, and history of concomitant illnesses such as asthma.

KEY POINTS.

• -Inhaled corticosteroids (ICS) should not be prescribed as monotherapy in COPD.

• -Combination therapy with ICS and long-acting beta-agonist (LABA) results in reduced exacerbations compared with bronchodilator treatment alone and triple therapy with ICS/LABA, and a long-acting muscarinic antagonist (LAMA) offers further benefit including a possible survival advantage in some patients.

• - In patients with higher blood eosinophil counts the use of ICS is associated with a greater reductions in COPD exacerbations risk.

• -ICS are associated with an increased risk of pneumonia. COPD patients with the highest pneumonia risk are those with low BMI, advanced age, male gender, greater airflow obstruction, and ICS containing treatments should be used with caution in these groups.

SYNOPSIS.

Inhaled corticosteroids (ICS), when used in combination with long-acting bronchodilators, reduce the risk of exacerbations and improve health-related quality of life in patients with COPD compared to bronchodilator or ICS therapy alone. Potential side effects of ICS include adverse effects on glycemic control, bone density, cataract formation, skin changes, oral candidiasis, and pulmonary infections. Pneumonia is observed at increased rates in COPD patients, particularly those with greater airflow limitation, low BMI, advanced age, and male gender and ICS may increase this risk further. Risk assessment is essential in selecting appropriate patients for ICS-containing therapy.

Box 1. Selecting ICS Candidates.

Consider ICS use	Avoid ICS use
Frequent and severe exacerbations (>2 moderate exacerbations or one hospitalization Eosinophilia Concomitant Asthma	Infrequent exacerbations/Stable Disease Frequent Pneumonia events Significant fractures in stable disease Low BMI <21