Intended for healthcare professionals

Clinical Review State of the Art Review

Asthma-COPD overlap syndrome: pathogenesis, clinical features, and therapeutic targets

BMJ 2017; 358 doi: (Published 25 September 2017) Cite this as: BMJ 2017;358:j3772
  1. Janice M Leung, assistant clinical professor of respiratory medicine,
  2. Don D Sin, professor of respiratory medicine
  1. Division of Respiratory Medicine and Centre for Heart Lung Innovation, St Paul’s Hospital, University of British Columbia, Vancouver, BC, Canada
  1. Correspondence to: D Sin don.sin{at}


Asthma-COPD overlap syndrome (ACOS) or asthma-COPD overlap captures the subset of patients with airways disease who have features of both asthma and chronic obstructive pulmonary disease (COPD). Although definitions of ACOS vary, it is generally thought to encompass persistent airflow limitation in a patient older than 40 years of age with either a history of asthma or large bronchodilator reversibility. ACOS affects about a quarter of patients with COPD and almost a third of patients who previously had asthma. Compared with their counterparts with asthma or COPD alone, patients with ACOS have significantly worse respiratory symptoms, poorer quality of life, and increased risk of exacerbations and hospital admissions. Whether this condition emerges after gradual shifts in airway remodelling and inflammation in a patient with COPD, as the result of noxious exposures in a patient with asthma, or even as a de novo disease with its own pathology is yet to be determined. Nevertheless, using treatments developed for asthma or COPD that target eosinophilic, neutrophilic, or paucigranulocytic airway inflammation may be a helpful approach to these patients until further clinical trials can be performed.


In 1961 at a bronchitis symposium in Groeningen, the Netherlands, Orie and Sluiter proposed what has since become known as the Dutch hypothesis, a paradigm under which asthma and chronic obstructive pulmonary disease (COPD) share a common origin with their divergence explained by an individual’s unique genetic make-up and environmental exposures.1 Despite the lasting influence of this paradigm, it was and continues to be vigorously contested,234 and, in reality, clinicians have practised for decades under the rubric of a fixed dichotomy between asthma and COPD. Only recently has a more fluid understanding of these two conditions re-emerged, existing instead along a continuum with a subset of patients showing features of both. Asthma-COPD overlap syndrome (ACOS) or asthma-COPD overlap is now thought to affect up to a quarter of patients previously thought to have COPD and up to a third of patients previously thought to have asthma.5678

In this review, we present the case for ACOS as a clinically important phenotype among airways diseases and advocate for the continued pursuit of its underlying pathophysiology. Our main focus is the potential mechanisms of disease, whether ACOS is merely an evolution from longstanding COPD or asthma or is its own disease entity with a unique pathogenesis. In particular, we evaluate the effects of smoking, air pollution, eosinophilic and neutrophilic mediated inflammation, proteolysis, genetics, and childhood lung development and how these factors might lead to ACOS. Although much remains to be determined about any of these factors in the development of ACOS, they provide a framework of understanding as to how features of asthma might develop in patients with COPD and vice versa. The latter part of the review will then use this framework to speculate on treatments that may benefit patients with ACOS and ultimately to establish priorities for the development of novel ACOS therapeutics.


The concept of ACOS is controversial, and even the use of the term “syndrome” may meet with some resistance as we do not yet know whether this is a single, unifying disease.9 We urge the medical community, however, to consider that many patients do not fall neatly within the categories of asthma and COPD and that many have a mix of features from both. Importantly, multiple studies have now shown that, compared with those with asthma or COPD alone, these patients face increased symptom burdens and higher rates of hospital admission and exacerbation,10111213141516 indicating that ACOS is a distinct phenotype in the airways disease spectrum and may have its own clinical trajectory. Unfortunately, treatments specifically targeted at this mixed phenotype are noticeably lacking. Patients with ACOS are often overlooked or simply excluded in clinical trials investigating novel treatments for asthma and COPD.1718 In addition, the paucity of data on the mechanistic drivers of ACOS and how these might differ from those of asthma or COPD has limited the discovery of ACOS specific treatments. As a result, how to identify and manage these patients remains a puzzle for clinicians. Investigation of this subset of patients with airways disease is imperative.

Sources and selection criteria

We obtained references through a PubMed search inclusive of publications from 2000 to March 2017. Search terms included “asthma COPD overlap syndrome”, “ACOS”, and “mixed asthma COPD phenotype”. We added additional references on the basis of the reference lists of recent review papers, including the 2016 consensus statement from Sin et al.17 We retrieved randomised controlled trials, observational cohort studies, systematic reviews, meta-analyses, and case series and reports for further review. We prioritised larger randomised controlled trials, observational cohort studies, and meta-analyses that were published most recently. We filtered search results for relevance to definition, epidemiology, mechanisms, and treatment. We considered only English language publications. For the section on emerging treatments, we searched for studies on ACOS, neutrophilic asthma, and eosinophilic COPD.

Definition and diagnosis

Given the nascence of ACOS, no single accepted definition for ACOS exists. Although it coined the term ACOS in its 2015 update on asthma management and prevention, the Global Initiative for Asthma (GINA) and the Global Initiative for Chronic Obstructive Lung Disease (GOLD) declined to put forth an overarching definition, citing the need for better phenotyping and further mechanistic work.19 More recently, the American Thoracic Society and the National Heart, Lung, and Blood Institute issued a workshop statement concluding that ACOS should not be considered a discrete disease entity, but rather an airways disease phenotype of mixed features.9 Nonetheless, GINA/GOLD identified diagnostic clues that suggest ACOS, including persistent yet reversible airflow limitation (post-bronchodilator forced expiratory volume in 1 s (FEV1) to forced vital capacity (FVC) ratio of <70% and FEV1 improvement of >12% and >400 mL from baseline after bronchodilator therapy); a history of asthma diagnosed by a doctor, atopy, allergies, or exposure to noxious agents; either sputum neutrophilia or eosinophilia; and age 40 years or older. Other groups, working largely through consensus opinion, have offered systems of major and minor criteria by which to identify patients who may have ACOS (fig 1).1720212223 Although subtleties exist between these proposed systems, they share several key obligatory features: that patients should be 40 years or older, have persistent airflow obstruction, and a history of asthma or evidence of bronchodilator reversibility. Curiously, many of these systems have not featured environmental exposures (either cigarette smoke or biomass fuel) as a key component despite the clear links between such exposures and the development of COPD. We propose that cigarette smoking or exposure to biomass fuel should be an integral component of the definition of ACOS, to avoid including patients with asthma who have developed airway remodelling and fixed airflow limitation. Given the lack of an accepted gold standard for the diagnosis of ACOS, none of the diagnostic schemes in fig 1 has been officially validated.


Fig 1 Consensus diagnostic criteria for asthma-COPD overlap syndrome (ACOS) in patients with airflow limitation. COPD=chronic obstructive pulmonary disease; CT=computed tomography; FeNO=fractional exhaled nitric oxide; FEV1=forced expiratory volume in 1 s; FVC=forced vital capacity; LLN=lower limit of normal

Of lesser importance in the diagnostic algorithm are biomarkers commonly used in the diagnosis and management of asthma. Although helpful in identifying COPD patients with possible asthma overlap, immunoglobulin E (IgE), sputum or peripheral blood eosinophil count, and fractional exhaled nitric oxide (FeNO) are not sufficiently sensitive or specific on their own for diagnosis of ACOS. Elevated IgE concentrations are found in up to half of patients with COPD,2425 and eosinophilic subtypes of COPD are well described in the literature.2627282930 The debate about which level of peripheral eosinophil count to ascribe to ACOS, whether above 2%, 3%, 5%, or instead using an absolute measurement, has also not been resolved.21313233 A FeNO cut-off of above 35 ppb may correlate with a post-bronchodilator FEV1 response greater than 200 mL and atopy in patients with COPD,2534 but the few studies that have looked at FeNO in ACOS seem to offer conflicting evidence of the test’s discriminatory power.635 Moreover, like sputum eosinophil measurements,3637 FeNO is neither widely available at all sites nor standardised across instruments or centres.3839 These factors remain adjunctive tools in the diagnosis of ACOS until further studies can be performed.

Epidemiology and clinical features

Without a standard definition or diagnostic system in place, the fact that population estimates of ACOS are highly variable comes as no surprise. Epidemiological studies have used definitions as disparate as the major-minor criteria system, self reported physician diagnoses, and combinations of spirometry results and symptom descriptions. Estimates of ACOS in the general population, using either self reported physician diagnoses or a combination of airflow obstruction on spirometry and symptom report, fall roughly between 2% and 3%.1240414243444546 In comparison, estimates of asthma and COPD in these same populations tend to be higher, around 5-17% for asthma and 2-12% for COPD. However, the prevalence of ACOS may vary according to geographic region, with estimates ranging anywhere from 0.61% in China to 3.7% in the United States.747 In cohorts of COPD patients specifically, the prevalence of ACOS ranges from 6% to 55%,6721464849505152535455565758 with pooled estimates from a meta-analysis suggesting a prevalence of just over 25%.5 In cohorts of asthma patients, the prevalence ranges from 10% to 31%.785559 In studies looking at patients with any airway obstruction (defined by either patient report or diagnosis in the medical record of asthma or COPD or through spirometry), the prevalence ranges from 15% to 56%.1113606162 Moving forward, comparisons of prevalence across regions and time will ultimately require a standardised definition for ACOS. Nevertheless, no matter which definition is used, ACOS seems to account for a considerable portion of all airways diseases.

Many case-control studies comparing ACOS with asthma and COPD suggest a specific demographic type: patients with ACOS tend to be younger,542474956576364 to be female,5663 and to have higher body mass index,54264 lower socioeconomic status,4142 and lower education levels.4142 ACOS patients also carry a higher burden of comorbidities, particularly for conditions such as gastro-oesophageal reflux disease, osteoarthritis, osteoporosis, depression, and anxiety.42445663 Importantly, despite divergent definitions and methods, these studies consistently show a distinct clinical trajectory for ACOS, one whose severity seems to exceed that of asthma or COPD alone. By various disease control measures—pulmonary function,15414345496566 respiratory symptoms,4349566567 exacerbation rates,144356586568 use of respiratory drugs,43565769 overall health status,74143 quality of life,51155 and disability124261—ACOS patients have greater decrements than their counterparts with asthma or COPD alone. Specifically, as high as a 20% decrement in per cent predicted FEV1 has been shown in ACOS patients compared with asthma patients,4143 and up to a 10% difference has been shown between ACOS and COPD patients.43 Exacerbation rates are up to four to five times higher in ACOS patients compared with asthma and COPD patients.4358 Emergency department visits,414261 hospital admissions,12131540424361 and healthcare use7861 are also significantly higher in ACOS patients, who incur twice the health related costs of asthma and COPD patients,7071 mainly through outpatient visits and drugs.72 The rate of physician visits, for example, is approximately 1.3 to 1.5 times higher in ACOS patients than in COPD patients.78

Data on excess decline in lung function (table 1) and excess mortality (table 2) in ACOS remain mixed. In a cohort of asthma, COPD, and ACOS patients, ACOS patients with late onset asthma (onset over the age of 40 years) had the greatest FEV1 decline (49.6 mL/year versus 27.3 mL/year in ACOS patients with early onset asthma and 39.5 mL/year in COPD patients).68 However, other studies have found that FEV1 decline in ACOS was similar to that in asthma but better than in COPD,40 or that no difference existed between ACOS, asthma, and COPD groups.62 Similarly, studies have shown increased mortality,454868 decreased mortality,1473 and no difference in mortality5556 when comparing ACOS with asthma and COPD. Discrepancies in case definition and variable follow-up times may account for these discordant results; in any case, the natural history of ACOS is still very much a question to be answered.

Table 1

Lung function decline in asthma-COPD overlap syndrome (ACOS)

View this table:
Table 2

Survival and mortality in asthma-COPD overlap syndrome (ACOS)

View this table:

Mechanisms of disease

The clinical presentation of ACOS in a manner that exceeds the expected disease burdens of either asthma or COPD alone raises an important question as to its origin: is ACOS the manifestation of a unique pathogenic mechanism, is it simply the additive (or synergistic) result of two distinct pathologies combined together, or is it the end result of various environmental stimuli and insults applied to a patient already affected by asthma or COPD? To answer this question, revisiting the principles of the Dutch hypothesis (and, conversely, its longstanding rival the British hypothesis) may be helpful. Fundamentally, the Dutch hypothesis posits that asthma and COPD are a single disease of diffuse airway obstruction (termed by Orie and Sluiter as “chronic non-specific lung disease”), but that both endogenous and exogenous factors help to steer patients into distinguishable phenotypes that we understand better today as asthma or COPD.74 These factors may include, but are certainly not limited to, genetics, sex, age, allergens, smoking, air pollution, biomass fuel, concomitant pulmonary diseases, and infections. On the other hand, the British hypothesis maintains separate origins for asthma and COPD, underscored by the fact that each disease has its own characteristic inflammatory profile, genetic polymorphisms, response to treatment, and clinical course.24

In this section, we consider ACOS within the prism of these two schools of thought (fig 2). If the Dutch hypothesis holds true, this would suggest that ACOS should take its place alongside asthma and COPD in the single disease continuum, emerging only through the confluence of specific patient related and environmental factors. Under this scheme, one could consider these three diseases as being somewhat in flux, that asthma leading to ACOS and COPD leading to ACOS are two potential pathways by which ACOS can develop. The British hypothesis as applied to ACOS would instead suggest that this is in fact its own disease with a unique genetic and molecular profile distinguishing it from asthma and COPD. Suffice to say, neither of these theories has yet been definitively proven or disproven when considering the conundrum of ACOS. We provide evidence that might support either side, but ultimately the message here is that further mechanistic work is desperately needed to unravel the pathogenesis of ACOS.


Fig 2 Possible mechanisms for the development of asthma-COPD overlap syndrome (ACOS). Two mechanisms by which ACOS might develop are presented. According to the Dutch hypothesis, asthma and chronic obstructive pulmonary disease (COPD) exist along a continuum with a shared origin. ACOS may fit along this continuum if, for instance, a patient with asthma is exposed to the usual inhalants that cause COPD, such as cigarette smoke or air pollution. In both humans with asthma and mouse models of asthma, these exposures lead to neutrophilic inflammation, cytokine release, oxidative stress, DNA methylation changes, and matrix metalloproteinase mediated proteolysis. Such changes may then shift the asthmatic airway to features of COPD, including fixed airways obstruction and emphysema. Conversely, the British hypothesis considers asthma and COPD to be unique diseases. Accordingly, ACOS may have its own origins in genetic predisposition or early childhood events. Whereas single nucleotide polymorphisms in CSMD1 and GPR65 show marginal genome-wide significance for ACOS compared with COPD alone, recent data suggest that attenuated lung development in early childhood may contribute to future development of ACOS

How might asthma become ACOS?

The gradual shift of a patient with asthma to one with features of both asthma and COPD requires several physiological adjustments to occur. Whereas asthma is conventionally viewed as a disease of variable airflow obstruction, often in response to allergen hypersensitivity,19 the obstruction in COPD is thought to be incompletely reversible and the product of environmental exposures such as tobacco or biomass fuel.75 Patients with asthma who are exposed to noxious inhalants traditionally associated with COPD could conceivably develop this degree of fixed obstruction over time. Secondly, in contrast to asthma, emphysema and the loss of lung elastic recoil is an important phenotype of COPD.76 Autopsy studies, however, suggest that these processes may have been overlooked in asthma.7778 Here, we review both aspects of COPD that may develop in patients with asthma.


A quarter of patients with asthma are current cigarette smokers,79 and in comparison with non-smoking patients with asthma, they carry a significantly greater risk for developing fixed airflow obstruction and ACOS.41808182 Exposure to cigarette smoke in asthma seems to remodel the airway in a way that may worsen small airways obstruction. Firstly, although goblet cell hyperplasia and mucus hypersecretion are well described pathological findings in both asthma and COPD, these features are aggravated in current smokers with asthma compared with ex-smokers or non-smokers with asthma.83 Secondly, multiple studies using both induced sputum samples and endobronchial biopsies have now shown that smoking in asthma increases airway neutrophilia, shifting the predominantly eosinophilic cell pattern typically observed in asthma towards a neutrophilic COPD pattern.8485868788 This process may be mediated by cytokines, as interleukin (IL)-6, IL-8, and IL-17A have all been implicated in neutrophil chemotaxis in smokers with asthma.85868789 In addition, smoking may induce CD8+ T cell proliferation in the asthmatic airway,90 similar to COPD.9192 Studies have shown that both neutrophil and CD8+ T cell numbers correlate well with increasing airflow obstruction.8593 Increases in CD8+ T cells in particular may relate to faster decline in lung function,93 as well as incomplete reversibility of airflow obstruction.94 Experimental models in which allergen sensitised mice are exposed to cigarette smoke seem to confirm these findings, with attenuation of eosinophilic inflammation in favour of increases in neutrophils, CD4+ and CD8+ T cells, and total cells.95969798

The recent application of new ’omics platforms to asthmatic smokers has shed light on novel genes and proteins that may play a role in shifting asthma to ACOS, although the direct connection from these findings to disease pathogenesis is still being clarified. Proteomic analysis of induced sputum samples in four different asthma phenotypes found that, compared with non-smoking patients with severe asthma, smokers and ex-smokers with severe asthma had decreased sputum abundance of tyrosine protein kinase Lyn (involved in mast cell degranulation).99 Genes up-regulated in the smoking group predominantly reflect actin cytoskeleton function. Validation of these target proteins and genes in independent cohorts and deeper interrogation of the mechanistic pathways underlying their role in smoking related changes in the asthmatic airway may provide further insight into the origins of ACOS.

Air pollution

Ambient air pollution—a mixture of particulate matter, carbon monoxide, lead, sulphur dioxide, nitrogen oxides, and ozone—is a well known mediator of airways disease, and multiple large population studies have shown that long term exposure to air pollution is detrimental for lung function. The Framingham Offspring or Third Generation studies, made up of more than 6000 participants, showed that proximity to a major roadway and exposure to fine particulate matter (PM2.5) decreased FEV1 and accelerated decline in FEV1.100 Other components of air pollution, such as nitrogen oxides and PM10, have also been associated with low FEV1.101 Conversely, improvements in PM10 and PM2.5 levels over time increased FEV1 and FVC in children residing in southern California.102 For patients with asthma in particular, exposure to excess air pollution has considerable deleterious effects, with exacerbations, emergency room visits, and hospital admissions all closely associated with levels of air pollution.103104105106107 The effect of air pollution on the asthmatic airway has multiple facets: impaired regulatory T cell function,108 activation of toll-like receptor (TLR)-2 and TLR-4 mediated innate immune responses,109 increased oxidative stress,110111112113 cytokine mediated airway inflammation,114115 and alterations in DNA methylation.116117

Any or all of these mechanisms could conceivably induce the phenotypic switch from asthma to ACOS, although data confirming this hypothesis are notably lacking. Nonetheless, a study investigating the development of COPD in Ontario’s asthma population provides the most convincing evidence to date for a link between air pollution and ACOS.8 In this study, 6040 adults in Ontario, Canada, with asthma were followed between 1996 and 2014, with a diagnosis of COPD by a physician during that time period used as the definition of ACOS. Air quality data were provided from 49 continuous fixed site monitors throughout Ontario. For every increase in PM2.5 of 10 μg/m3, the adjusted hazard ratio for developing ACOS was 2.78 (95% confidence interval 1.62 to 4.78). Environmental exposures may thus be one important trigger in the pathogenesis of ACOS.

Loss of lung elastic recoil and development of emphysema in asthma

Emphysema, hyperinflation, and the loss of lung elastic recoil are often associated with COPD, but they may also play an integral role in the later stages of severe asthma. Recognition of the loss of lung elastic recoil dates back to 1967 when Gold, Kaufman, and Nadel showed that patients with severe asthma lasting longer than five years had significantly reduced lung elastic recoil compared with healthy people.118 Initially, this was not thought to be related to emphysematous tissue destruction, as elastic recoil would return to normal following treatment with bronchodilators. However, subsequent studies showed that the loss of lung elastic recoil was more durable than anticipated and that this loss can still be observed in stable clinical states.119120121 Normal diffusing capacity and computed tomography imaging again argued against the presence of significant emphysema as a cause of elastic recoil abnormalities.120

Two autopsy studies have since shifted our understanding of the structural lung changes in asthma. A study in 2004 found a decrease in elastic fibres and an increase in the number of abnormal alveolar attachments in fatal asthma compared with controls.78 A subsequent study in 2014 reported the autopsy findings of three patients with asthma, which showed that although macroscopic evaluation of the lungs did not show emphysema, microscopic examination showed diffuse upper lobe predominant mild centrilobular emphysema.122 More recently, it has been suggested that recurrent exacerbations of asthma induce bronchiolar inflammation, with activation of proteases such as neutrophil elastase, cathepsin G, and matrix metalloproteases (MMP), all of which conspire to break down the lung parenchyma.77 In vitro work has implicated IL-13 as a potent suppressor of elastin mRNA expression through a pathway that involves the secretion of MMP-1 and MMP-2.123 The later stages of asthma may thus start to resemble emphysema, shifting asthma down the airways spectrum to ACOS.

How might COPD become ACOS?

The reverse scenario, in which a patient with COPD develops features of asthma, is admittedly less compelling. This would require the development of three important pathological traits of asthma in a patient with COPD: allergen sensitisation, airway hyper-responsiveness, and eosinophilic and type 2 (Th2) mediated airway inflammation. Allergen sensitisation has been known to occur in COPD, particularly in older people,124 and upwards of 25-30% of COPD patients report allergic upper airway symptoms or show IgE sensitisation to perennial allergens.125 Among smokers, allergen sensitisation has also been associated with faster decline in lung function and, among COPD patients, with increased respiratory symptoms and COPD exacerbations.125126 Whether allergen sensitisation ultimately remodels the airway in the manner of asthma has yet to be demonstrated, however, and little research on this has been done. Similarly, airway hyper-responsiveness and eosinophilic/Th2 mediated inflammation also occur in COPD. However, for the former it is by no means certain whether the pathological process is the same as in asthma, and for the latter it is conceivable that similar inflammatory changes also occur in patients with asthma who develop ACOS. Below, we review the mechanisms by which these two features may emerge or worsen over the course of COPD.

Airway hyper-responsiveness and airway remodelling

Airway hyper-responsiveness is the acute response of an airway to a bronchial challenge, often in the form of an agent such as methacholine or histamine. Airway hyper-responsiveness affected up to a quarter of participants in the Lung Health Study and was associated with faster decline in FEV1 and higher respiratory related mortality.48 The response to a bronchial challenge is approximately 25-50% more robust in moderate to severe COPD than in mild COPD. This might suggest that this later stage development may simply be geometric in nature, as the resistance of the airway is inversely proportional to the radius to the fourth power (and therefore may merely represent advanced airflow obstruction regardless of the underlying airways disease).127128 Airway wall thickening, airway smooth muscle (ASM) proliferation, and the obliteration of the terminal bronchioles develop over time in COPD and, along with mucous secretions, contribute to this airway narrowing.129130131 The degree of airway hyper-responsiveness depends on the balance between ASM shortening and the opposite recoil force of the surrounding tissue, which in COPD may be compromised by the development of emphysema and the destruction of elastic fibres.132 As the airway wall thickens and as ASM area increases, a larger ASM contraction against a weaker opposing elastic load allows for a larger degree of airway narrowing. Whether we can easily equate the process of airway hyper-responsiveness in COPD with that in asthma is not clear, however. The contribution of changes in the extracellular matrix to airway hyper-responsiveness is much larger in COPD than in asthma, whereas the airway hyper-responsiveness in asthma is driven primarily by ASM hypertrophy and hyperplasia.133 Therefore, although these processes may look similar on provocation testing, treatments targeting airway hyper-responsiveness in asthma may not necessarily translate to COPD patients who develop airway hyper-responsiveness and evolve to ACOS.

Airway inflammation

Conventional understanding of COPD places emphasis on a neutrophilic, type 1 (Th1) inflammatory airway response,134135 but a subset of COPD patients who do not cleanly fit this paradigm clearly exist. For instance, peripheral blood eosinophil concentrations persistently above 2% over a three year period were found in just under 40% of participants in the ECLIPSE (Evaluation of COPD Longitudinally to Identify Predictive Surrogate End-points) cohort.136 Recently, Christensen et al applied a 100 gene signature corresponding to high Th2 inflammation to airway epithelial cells from COPD patients.137 Approximately 5% of smokers with COPD showed a high Th2 gene expression signature, which was associated with increased serum and tissue eosinophil counts, increased bronchodilator responsiveness, and greater improvement in hyperinflation after treatment with inhaled corticosteroids. This signature also notably correlated with the gene expression of mast cell marker CPA3 and eosinophil chemotactic molecule CCL26. Whether this signature has identified a unique endotype of COPD or has isolated an important ACOS airway epithelial gene signature is unclear; the authors recommend that the signature be studied in cohorts better phenotyped for both these conditions.

Despite the concerns about the quality of phenotyping in Christensen et al’s cohorts, other authors investigating sputum cytokine clusters would concur with their hypothesis that an overlap phenotype of eosinophilic COPD can be readily recognised.138 In a cohort of patients with severe asthma and COPD, three clusters were identified on the basis of sputum cytokine measurements: asthma with high Th2 and eosinophilic inflammation, asthma and COPD with sputum neutrophil predominance, and COPD with predominantly eosinophilic inflammation. This latter group, corresponding to the high Th2 COPD group in Christenson et al’s study, featured increased concentrations of important Th2 cytokines and chemokines—namely, IL-6, CCL2, CCL13, and CCL17. Although these studies have been instrumental in exposing the significance of Th2 responses in COPD, why certain COPD patients are driven towards this response compared with others who remain in the Th1, neutrophilic realm is still unclear.

Is ACOS its own disease?

Finally, in accordance with the British hypothesis, we consider a third scenario in which ACOS has a unique origin unrelated to either asthma or COPD. Some of the strongest evidence against the Dutch hypothesis has been the striking lack of a unifying genetic cause of asthma and COPD despite many genome-wide association studies (GWAS).139 Whether ACOS will follow this pattern and show its own genetic determinants separate from those of asthma or COPD has yet to be definitively answered. One GWAS comparing patients with COPD and ACOS in the COPDGene Study did not find any single nucleotide polymorphisms (SNPs) associated with ACOS that met the pre-determined genome-wide significance threshold of 5×10−8, although the authors noted that the study was probably underpowered to detect this degree of difference.16 SNPs that approached genome-wide significance included those on the genes CSMD1 and GPR65, the former a tumour suppressor gene previously implicated in emphysema and the latter a mediator of eosinophil activation in asthma.140141 Clearly, genetic work on ACOS remains preliminary and larger studies are needed to better establish its genetic determinants.

If asthma and COPD are not in fact prerequisite stages to be passed through before developing ACOS, it is conceivable that events or insults in early childhood might trigger the shift towards asthma, COPD, or ACOS, each having its own distinct trajectory. In support of this hypothesis, the Tasmanian Longitudinal Health Study analysed lung function collected at age 7 and then at age 45, determining whether childhood spirometry could predict the development of ACOS in adulthood (defined as a self report of asthma in conjunction with a post-bronchodilator FEV1/FVC below the lower limit of normal).46 Those in the lowest quarters of FEV1/FVC and FEV1 per cent predicted at age 7 had a 16.3 and 2.93 greater odds, respectively, of developing ACOS by age 45. This association, which persisted even after accounting for smoking, was far greater than that between childhood lung function and COPD, whereas no association with asthma was noted. Whether due to in utero and/or childhood exposure to smoke or childhood respiratory infections, the inability to reach maximum lung function during childhood, already discernible by age 7, seems to have lasting implications for future airways disease.

Possibilities for therapeutic intervention

The myriad ways in which a particular person can potentially develop ACOS poses considerable challenges for treatment. The very heterogeneity of ACOS, the untoward by-product of a lack of consensus surrounding its pathophysiology and even its definition, implies that no single class of treatment is likely to help all patients. Importantly, we and others have advocated that the field move towards deeper phenotyping of these patients to guide treatment. In particular, as proposed by Barnes, three phenotypes of ACOS have been described that are important to establish before starting treatment: eosinophilic COPD (high Th2 inflammation), neutrophilic asthma (low Th2 inflammation), and paucigranulocytic ACOS.142 Needless to say, improved access to and standardisation of sputum cytology measurements is needed before such a strategy can be implemented. Even so, this framework provides a useful algorithm for targeting treatment of ACOS, one rooted in the presumed pathophysiology of the disease. In this section, we present the evidence for therapeutic options based on these three phenotypes, with the caveat that few trials have specifically investigated patients with ACOS. Extrapolations from asthma and COPD trials are thus also presented.

Eosinophilic COPD

Inhaled corticosteroids

Targeted against eosinophilic pathology, inhaled corticosteroids and combinations with long acting β agonists (LABA) have naturally featured prominently in treatment recommendations for ACOS given the emphasis many groups have placed on eosinophilia as a diagnostic criterion.143144 In addition, inhaled corticosteroid/LABA combinations are recommended in both asthma and COPD, making their use in ACOS appealing. The strength of evidence for these recommendations is mixed, however. Trials evaluating inhaled corticosteroids alone and in combination with LABA have been small in size and mostly retrospective and observational in nature. Moreover, results have been conflicting. A 12 year observational retrospective cohort study in which 90 patients with ACOS treated with inhaled corticosteroids were compared with 35 ACOS patients not treated with inhaled corticosteroids found no difference in decline in FEV1, exacerbation rates, or overall mortality.145 On the other hand, a prospective study found that 127 ACOS patients who received inhaled budesonide had improved spirometry and measures of hyperinflation, sputum eosinophils, serum IgE, and FeNO after treatment.146 The lack of a true placebo group in this trial, however, limits the generalisability of these results. Compared with COPD patients, those with ACOS may show better FEV1 response to inhaled corticosteroid/LABA treatments, but again, these trials featured no more than 45 ACOS patients.52147 The possibility that inhaled corticosteroids with or without LABA may provide benefit for eosinophilic ACOS patients cannot be either discounted or definitively proven given the amount and quality of data at hand.

Anti-IgE therapies

Severe, persistent asthma that remains uncontrolled despite inhaled corticosteroids and inhaled corticosteroid/LABA combinations now prompts consideration of anti-IgE treatments, including recombinant humanised monoclonal antibodies such as omalizumab. The binding of IgE to the IgE receptor FcεR1 on basophils and mast cells triggers cytokine mediated eosinophilic inflammation; blockage of this linkage by compounds such as omalizumab helps to attenuate this cascade, which ultimately has been shown to reduce exacerbations, hospital admissions, emergency department visits, and inhaled corticosteroid and rescue inhaler use in patients with advanced asthma.148 Recently, the efficacy of omalizumab was assessed in patients with severe allergic asthma enrolled in the Australian Xolair Registry who also had a diagnosis of COPD, whether by physician assessment (n=17) or by fixed airflow obstruction on spirometry (n=55).149 Regardless of the definition of COPD, these ACOS patients showed significantly improved asthma control and health related quality of life scores (exceeding the minimally clinically important difference) following treatment with omalizumab despite no significant improvements in their FEV1. These incremental gains were equivalent to those observed in patients with asthma alone. Smaller case series of three and 10 ACOS patients treated with omalizumab each echoed these improvements in asthma symptom control.150151

Anti-IL-5/IL-5Rα therapies

IL-5 plays a key role in eosinophil differentiation and maturation in the bone marrow, as well as migration from the blood to tissue sites.152 Three new anti-IL-5 and anti-IL-5Rα monoclonal antibodies (mepolizumab, benralizumab, and reslizumab) have now been studied in eosinophilic asthma with impressive improvements in exacerbation rates, asthma control scores, and steroid usage.153154155156 These compounds have yet to be tested in cohorts of ACOS patients, but a randomised, double blind, placebo controlled study evaluating benralizumab in COPD patients with sputum eosinophil counts above 3% was far from promising.157 Benralizumab failed to reduce the exacerbation rate, the primary endpoint, although improvements in spirometry seemed to occur. For the subgroup of patients with baseline blood eosinophil counts above 200 cells/μL, numerical but non-significant improvements in exacerbation rates, FEV1, and St George’s Respiratory and Chronic Respiratory Disease Questionnaire scores occurred, but until further large randomised controlled trials are conducted, the role for anti-IL-5 and anti-IL-5Rα monoclonal antibodies in eosinophilic COPD is unclear.

Anti-IL-13/IL-4Rα therapies

IL-13 and IL-4 are important Th2 cytokines that contribute to several asthma related pathologies including the regulation of eosinophils, mucus production, goblet cell hyperplasia, ASM contraction, and airway remodelling.148 Anti-IL-13 antibodies such as lebrikizumab and tralokinumab target IL-13, whereas anti-IL-4Rα antibodies such as dupilumab target both IL-13 and IL-4. In patients with severe, persistent asthma, these monoclonal antibodies have shown improvements in lung function and exacerbation rates.158159160161162 The efficacy of these agents seems to be enhanced in patients with high serum concentrations of periostin, an extracellular matrix protein deposited in the basement membrane of asthmatic airways that reflects Th2 inflammation.163 Although the IL-4Rα antagonists seem to show greater efficacy against asthma than do the IL-13 agents, both have yet to be tested in eosinophilic COPD, so extrapolation to ACOS is merely speculative at this time.

Neutrophilic asthma—macrolides

The pleiotropic properties of macrolides, functioning at once as antibacterial, immunomodulatory, and anti-inflammatory drugs, have resulted in their wide application in several neutrophilic respiratory disorders such as cystic fibrosis and non-cystic fibrosis bronchiectasis. Macrolides seem to dampen IL-8 and CXCL1, leading to neutrophil apoptosis and decreased oxidative stress.164 Importantly, azithromycin has become a therapeutic option in COPD, in which its long term use was shown to reduce the frequency of exacerbations,165 and whether these effects could translate to neutrophilic forms of asthma has been the subject of several studies. To date, the results have been modest at best. Although the larger randomised controlled trials focusing on patients with non-eosinophilic asthma have shown a reduction in exacerbation rates and duration of acute events with long term macrolide use,166167 other trials have shown no significant benefit on either these outcomes or asthma symptom control and lung function.168169170171172173174 The 2015 Cochrane review of macrolides in chronic asthma concluded that, on the basis of very low quality evidence, macrolides had no discernible benefit in reducing exacerbations or in improving asthma control, quality of life, and rescue medication use.175 A small positive effect on lung function could be observed, suggesting that there may still be some clinical benefit to certain asthma patients; however, large effect sizes for clinically important endpoints in asthma are unlikely to be achieved with macrolides.

Paucigranulocytic ACOS—long acting muscarinic antagonists

For those ACOS patients without a distinct cellular inflammatory response, we return to bronchodilators, the stalwart of airways disease treatment. Ideally, for the patient with ACOS, bronchodilators should reduce ASM tone and dynamic hyperinflation. We have described the role of inhaled corticosteroid/LABA combinations in ACOS above; here we discuss whether long acting muscarinic antagonists (LAMA) widely used in COPD have any benefit in asthma and, by extension, ACOS. Studies assessing the efficacy of LAMA therapy in asthma have largely evaluated tiotropium, an M1 and M3 receptor antagonist with a proven benefit in COPD. Several large, high quality randomised controlled trials (one of which enrolled asthma patients with persistent post-bronchodilator airflow limitation176) have consistently shown improvements in FEV1 in patients with asthma, both in children and in adults, when tiotropium is added to inhaled corticosteroid and inhaled corticosteroid/LABA therapies.176177178179180181182 Several studies have also shown a reduction in exacerbation risk,179183 as well as asthma symptom control.179181182 Whether LAMA compounds other than tiotropium have similar positive effects is yet to be tested, but the clinical improvements observed in both COPD and asthma patients suggest that LAMA therapies may be an option for patients with ACOS.

Emerging treatments

Although several ongoing studies are aiming to characterise ACOS further, no clinical trials are being conducted to assess the utility of either novel or known treatments in ACOS patients. Furthermore, attempts to apply available asthma therapies to COPD patients with an eosinophilic phenotype have garnered mixed results. One phase III trial evaluating mepolizumab in eosinophilic COPD showed a statistically significant reduction in moderate to severe exacerbations compared with placebo, whereas another failed to show any difference.184 Whether future trials will be performed to resolve these conflicting results has yet to be announced.


To date, no guideline statements grading the quality of evidence surrounding ACOS have been issued. Expert panel consensus statements (fig 1), however, have been published with general agreement about two principles: that ACOS should be considered in any patient over the age of 40 with persistent airflow obstruction and a history of asthma or evidence of bronchodilator reversibility and that the treatment of ACOS should include inhaled corticosteroids. However, these statements acknowledge the need for additional evidence for these recommendations.


ACOS may represent a distinct clinical phenotype of asthma and COPD and, importantly for clinicians, is marked by worse respiratory symptoms and rates of exacerbation and hospital admission. For these reasons, the diagnosis of ACOS should be considered in all patients over the age of 40 years with persistent airflow obstruction, a history of cigarette smoking or exposure to biomass fuel, and a history of asthma or strong bronchodilator reversibility on spirometry. Although the pathway to ACOS has yet to be elucidated in its entirety, one can speculate that several ways exist in which a patient with asthma can develop a neutrophilic “COPD-like” airway and a COPD patient can develop an eosinophilic, reactive “asthma-like” airway. Basing therapeutic choices on these inflammatory distinctions may help to guide the clinician in an otherwise data-free zone. Greater emphasis on the ACOS population in future research is recommended.

Glossary of abbreviations

ACOS—asthma-COPD overlap syndrome

ASM—airway smooth muscle

COPD—chronic obstructive pulmonary disease

FeNO—fractional exhaled nitric oxide

FEV1—forced expiratory volume in 1 s

FVC—forced vital capacity

GINA—Global Initiative for Asthma

GOLD—Global Initiative for Chronic Obstructive Lung Disease

GWAS—genome-wide association studies

IgE—immunoglobulin E


LABA—long acting β agonist

LAMA—long acting muscarinic antagonist

MMP—matrix metalloprotease

SNPs—single nucleotide polymorphism

Questions for future research

  • Do unique genetic, molecular, and histological patterns exist that distinguish asthma-COPD overlap syndrome (ACOS) from asthma or chronic obstructive pulmonary disease (COPD) alone?

  • What are the long term clinical outcomes for patients with ACOS, and how might these outcomes vary depending on the definition used for ACOS?

  • Should first line treatment for patients with ACOS include inhalers containing inhaled corticosteroid?

  • Do patients with ACOS benefit from standard treatments for asthma and COPD, or are novel compounds needed specifically tailored to the ACOS phenotype?


  • Series explanation: State of the Art Reviews are commissioned on the basis of their relevance to academics and specialists in the US and internationally. For this reason they are written predominantly by US authors

  • We thank Cheng Wei Tony Yang and Chloe Chih Ya Huang for their contribution to the figures.

  • Contributors: Both authors were involved in the conception, writing, and editing of the manuscript. Both are guarantors.

  • Competing interests: We have read and understood BMJ policy on declaration of interests and declare the following interests: DDS has received honorariums for speaking engagements with AstraZeneca and Boehringer Ingelheim and for participating in COPD advisory boards of Sanofi, Regeneron, AstraZeneca, Boehringer Ingelheim, Novartis, and Merck and has received research funding for investigator initiated COPD studies from Merck, AstraZeneca, and Boehringer Ingelheim. DDS is supported by a Tier 1 Canada research chair award in COPD.

  • Provenance and peer review: Commissioned; externally peer reviewed.

  • Patient involvement: No patients were involved in the creation of this article.


View Abstract