Chronic obstructive pulmonary disease (COPD) has classically encompassed the pathological entities of emphysema and chronic bronchitis. However, in the last decade these terms are no longer included in the formal definition; instead, national and international guidelines have defined it spirometrically,^1,2 to the extent that the pathology receives scant attention. Doctors have progressively been advised to manage patients exclusively through spirometry, symptoms and exacerbations with no consideration of the underlying disease process.³ Indeed, in the most recent guideline, the pathology is not mentioned at all.³

The best practice guidelines recommend that:^1,2,3

Herein, the above claims are dissected via reviewing the data on BD alone as the foundation of treatment and the apparently minor role of ICS. The validity of evidence regarding pneumonia and the issue of steroid withdrawal will also be discussed.

All the data that will be reviewed are derived only from smoking-related COPD. Other important aspects such as smoking cessation and rehabilitation will not be expounded.

It must be the supreme paradox. Chronic obstructive pulmonary disease is defined as a spirometrically irreversible disease, and subjects are then given BD to see how much the lung function can improve. Chronic obstructive pulmonary disease as a disease is not being understood. It is not asthma, where smooth muscle hyperplasia and excessive bronchoconstriction occur and which is very responsive to BD. This is not the bronchial pathology of COPD, which manifests as irreversible structural changes with persistent airway narrowing, inflammation and mucus gland hypersecretion. One must therefore appreciate that monitoring lung function serially to assess response to treatment is actually severely limited. This is one of the reasons why the effect of ICS cannot be appreciated – it cannot be monitored appreciably with spirometry. With remodelling and a somewhat fixed airflow limitation, even the most potent anti-inflammatory agent will not result in substantial changes in lung function. Despite these constraints, up to 60% of COPD subjects may show spirometric reversibility as defined by the American Thoracic Society.^4,5

The Global Initiative for Obstructive Lung Disease (GOLD) and South African guidelines consider the definition for COPD exemplified by the post-bronchodilator forced expiratory volume in one second (FEV₁)/forced vital capacity (FVC) of < 0.70 as being diagnostic.^1,2 Another limitation, however, is that the criterion is neither sensitive nor definitive for the detection of COPD.⁴ The SPIROMICS (Subpopulations and Intermediate Outcome Measures in COPD Study) research group reported on an observational study involving 2736 current or former smokers.⁶ Respiratory symptoms were present in 50% of subjects with spirometry in the normal range (who would not have met the COPD definition). These participants had:

These data suggest that we need to be circumspect about the GOLD definition, carefully evaluate at-risk patients for symptoms and treat them appropriately.

A disconnect also exists between spirometry and the severity of COPD. Even in mild disease, as defined by FEV₁ of 80% or greater of the predicted value, most transitional bronchioles and about a third of the alveolar surface of the lung have already been destroyed.⁷

Inflammation is present in the early stages of disease, is known to be neutrophil-driven and leads to chronic bronchitis and damage to the lung parenchyma. Inhaled corticosteroids are known not to have a major antineutrophil effect.⁸ Early studies on the anti-inflammatory actions of ICS in COPD demonstrated that they were not as rapid as in asthmatics, dissuading clinicians. However, when these were studied over extended periods, multiple anti-inflammatory effects were noted.⁹ One reason for the limited responsiveness to ICS has also been ascribed to reduced histone deacetylase 2 (HDAC2) expression.¹⁰ It was subsequently found that certain anti-inflammatory effects of steroids are independent of HDAC2.¹¹ Thus, this pathway can be bypassed for ICS to ameliorate tissue damage. A meta-analysis of studies that evaluated inflammatory biomarkers in sputum, bronchoalveolar lavage fluid and biopsy specimens demonstrated that ICS led to a reduction in cluster of differentiation 4 (CD4+), CD8+, neutrophils and lymphocytes.¹² Another study concluded that even with persistent smoking, long-term ICS treatment could still produce anti-inflammatory effects in the lung.¹³

A Dutch group has done extensive studies of ICS in COPD.^14,15 In one, with fluticasone propionate (FP) and salmeterol, almost 300 genes involved in the pathogenesis exhibited altered expression after treatment.¹⁵ Clinically, the FEV₁ in every patient improved pari passu with the ICS, but this ranged from poor to good. As in all studies, this Gaussian gradation of responses reflects the variability of steroid responsiveness of the pathogenetic mechanisms and the capacity for bronchodilator reversibility from the remodelling process.

The synergy between molecular interactions of ICS and long-acting beta agonists (LABA) are well recognised in asthma and occur in COPD as well; certain in vitro effects of high-dose ICS could be achieved with lower doses of ICS when combined with LABA.¹⁶

Clinically, the addition of ICS to BD results in a relative improvement in lung function, quality of life and exacerbation reduction – all to the order of approximately 25%.¹⁷ Exacerbations are particularly important. These can be life threatening, are major cost-drivers and can result in acute permanent deterioration in the trajectory of functional decline. Hence, this is a major end point in clinical studies, and it is only supplementary ICS that influence this substantially. Because exacerbations tend to occur in those with prior such episodes and more severe disease (FEV₁ < 60%), this group has been the focus of study. However, even those without such an experience and FEV₁ of > 60% may also benefit from ICS. This was observed in the Study to Understand Mortality and Morbidity in COPD (SUMMIT) sub-study and the Efficacy and Safety of Triple Therapy in Obstructive Lung Disease (ETHOS) clinical trial.^18,19

The only other agent to address inflammation that has shown potential is the phosphodiesterase-4 inhibitor roflumilast. It has demonstrated a reduction in COPD exacerbations with improvement in the quality of life (QOL) and lung function.¹ It is the final choice in the GOLD therapeutic algorithm, indicated only in chronic bronchitis, but some side effects may limit its use.

For the COPD participants, having higher eosinophil counts in blood or sputum – an ‘eosinophilic phenotype’ – was noted to confer ICS responsiveness.^29,30 It is now generally accepted that COPD patients with eosinophilia (≥ 2% eosinophils or ≥ 150 eosinophils/µL) respond better to ICS therapy. A post hoc analysis of the Investigating New Standards for Prophylaxis in Reduction of Exacerbations study (INSPIRE) – which utilised a 2% eosinophil cut-off – demonstrated that FP/salmeterol was linked with a 25% exacerbation risk reduction compared to tiotropium alone.³¹

In the WISDOM study, dual long-acting BD and maintenance FP exhibited a linear relationship of further exacerbation reduction with higher eosinophil counts compared to dual BD alone – in other words, the higher the eosinophil count, the better the exacerbation reduction.³² In another, the Study to Understand the Safety and Efficacy of ICS Withdrawal from Triple Therapy in COPD (SUNSET), patients with high eosinophils went on to experience the largest lung function loss and exacerbation risk, prompting the authors to state that in this group, ICS should not be discontinued.²³

The progressive nature of COPD, intractable symptoms and the increasing propensity to exacerbations mean that there is a need to sequentially add treatment. The early options were a long-acting beta agonist (LABA), and/or a long-acting muscarinic antagonist (LAMA), as these are superior to their short-acting counterparts, and ICS in various combinations and devices. An ICS/LABA is superior to LAMA monotherapy.³³ The need for all three agents and short-acting agents for rescue could mean up to four different inhalers. Such polypharmacy does not aid adherence. Moreover, COPD and age are accompanied by increasing frailty, difficulty in inhalation techniques and coordination. As a proof of concept, multiple inhaler triple therapy was investigated. In the Lung Function and Quality of Life Assessment in COPD with Closed Triple Therapy (FULFIL) trial, umeclidinium (Umec) was added to FF/VI and resulted in a significant approximately 180-millilitres (mL) FEV₁ net gain and QOL improvement when compared to budesonide (Bud)/FF at the one year mark.³⁴ Single inhaler triple therapy (SITT) was then embarked upon, with all consistently showing improvements in lung function, QOL measures and exacerbation reduction. In TRIBUTE, a 15% reduction in exacerbation rate was noted with extra-fine beclomethasone dipropionate, formoterol furoate and glycopyrronium (G) bromide (BDP/Form/GB) compared with indacaterol/GB.²⁸ Higher rate reductions were demonstrated in the other SITT studies, IMPACT (FF/Vi/Umec) and ETHOS (Bud/Form/G) (Table 2).^23,35

A mortality benefit is the most sought-after goal in COPD. It was hoped that an agent that was potentially ‘disease modifying’, such as ICS, would prove beneficial over BD. An early study showed a numerical benefit, but this did not reach statistical significance (TORCH).²¹ This threshold was finally passed with the IMPACT study, when the ICS (FF), either with dual BD (umeclidinium and vilanterol) or vilanterol (FF/VI), proved superior to the dual BD.³⁵ This is likely a result of FF – one of the most recent glucocorticoids to be developed – which has the highest receptor-binding properties of all ICS with pronounced anti-inflammatory effects and a reduction in exacerbations. The only other trial to show a mortality benefit was ETHOS, utilising high-dose budesonide and glycopyrronium–formoterol.²³

There is an emerging strategy that when the disease is stable, then the ICS component in either a dual or triple treatment regimen, because of the perceived risks, can be discontinued³⁶; this is strongly discouraged. It also stems from a parallel in asthma where a reduced ICS dose is required when the inflammation subsides. This is not true of COPD, as the ICS is maintaining stability of the disease. In the development of COPD medication, dual therapy including ICS was first introduced and became the standard of care. With the caution around ICS, studies were then performed where the ICS was withdrawn. All consistently showed that that this was deleterious; patients rapidly experienced exacerbations,^37,38 a persistent decrease in lung function³⁹ and lowered quality of life.³⁷ As anticholinergics became available, these were progressively added to the ICS + LABA: triple therapy. An attempt was again made to omit the ICS. A salutary landmark study showed that this was an unwise decision. In the WISDOM study, subjects on tiotropium, salmeterol and fluticasone propionate (FP, 1000 microgram [µg]/d) gradually had the FP withdrawn. In those maintained on dual BD only, the FEV₁ immediately and progressive declined over one year, as did the QOL, and was not equivalent in exacerbation protection. In the SUNSET study, lung function deteriorated in those maintained on indacaterol–glycopyrronium only compared to those on tiotropium and salmeterol–FP.

A second major reason not to discontinue ICS is a failure to appreciate the natural history of COPD. The clinical course of COPD is characterised by accelerated decline in lung function and concomitant functional capacity, as was elegantly demonstrated by the famous graph of Fletcher and Peto.⁴⁰ Because of this trajectory, therapy needs to be augmented progressively in COPD rather than withdrawn. In general, more treatment is required to improve symptoms and prevent exacerbations with the passage of time. In a general practice report, 30% of newly diagnosed patients in primary care had to progress to triple therapy (dual BD + ICS) within two years of presentation.⁴¹

The augmentation of clinical improvements and course of the disease with ICS when added to mono or dual BD clearly attest to the importance of suppressing inflammation and indicate that this component should not be forgotten. This effect is particularly pronounced in exacerbation reduction. The hesitancy in prescribing ICS because of the possibility of adverse metabolic effects and pneumonia, although important, is unwarranted. When confounders for these occurrences have been considered, the risks appear negligible or of very low frequency at worst. The data also show that ICS withdrawal is deleterious and ill-advised in a disease with a natural history of progressive decline necessitating increased pharmaceutical support. Those patients with more eosinophils should be targeted for early ICS therapy, and de-escalation should not be contemplated for them. The pharmacotherapy discussed is available in several formulations singly and in combination, with varying extent of availability and cost in different countries. The clinician should be prudent in choice, preferably choosing an ICS in the regimen. When the disease has progressed beyond the mild stage, dual BD and ICS, ideally in a single inhaler, are a compelling option for optimal therapeutic outcomes.