The heterogeneity of COPD exacerbations reflects their dependence on a complex spectrum of interacting factors. Susceptibility to exacerbations and the basis of the frequent exacerbator is therefore probably multifactorial. Identifying contributory factors, such as background airway inflammation, airway microbial patterns and host immunological responses, may provide potential targets in the effort to alter a patient’s exacerbation frequency phenotype.
Airway and systemic inflammation at exacerbation and in the frequent exacerbator
During exacerbations of COPD the existing airway and systemic inflammation associated with this inflammatory disease increases further [9–14]. An inflammatory cascade releases inflammatory mediators, such as interleukins and chemokines, which recruit and activate immune cells. This cascade is thought to contribute to the local structural damage responsible for COPD progression, but also causes systemic inflammation with increases in acute phase proteins such as fibrinogen  and C reactive protein (CRP) . Although the exacerbation inflammatory process is predominantly neutrophilic [16, 17], the numbers of all inflammatory cell types within bronchial mucosa increase and the inflammatory character is also influenced by the type of insult triggering an exacerbation. For example, compared to bacterially driven exacerbations viruses appear to stimulate more eosinophil activity [17, 18].
The airways of frequent exacerbators, however, appear to be more inflamed with higher levels of sputum interleukin (IL)-6 and IL-8 even in the stable state . In addition, their trajectory of inflammation is also worse with sputum IL-6 and plasma fibrinogen  continuing to increase more rapidly over time.
Frequent exacerbators have higher sputum IL-6 and serum CRP during their exacerbation recovery periods , making persistent post-exacerbation inflammation a possible explanation for their higher baseline inflammation. A higher CRP during the recovery period is also associated with a shorter time until the next exacerbation  perhaps suggesting an inflammatory basis to the temporal clustering of exacerbations reported by Hurst et al., who identified the 8-week post-exacerbation period as particularly high risk for exacerbation recurrence .
Persisting and increased propensity to inflammation therefore is a feature of the frequent exacerbator phenotype, although whether these features are causative or are downstream of susceptibility to the triggers of exacerbations remains to be established.
Small airway obstruction in COPD is accompanied by dynamic lung hyperinflation and this is a major driver of respiratory symptoms, particularly dyspnea . At exacerbation, airway inflammation and worsened expiratory flow limitation worsen this, leading to further dynamic hyperinflation with increased respiratory effort, cardiovascular strain, inspiratory muscle overload and potential respiratory failure . In those patients with severe airflow obstruction and a lower FEV1% predicted, hyperinflation is more likely to be present even in stable disease and there may be reduced capacity to deal with any further insult. Exacerbations would potentially be more easily triggered in these patients and this may explain why a higher exacerbation frequency is seen with more severe airflow restriction.
Bronchodilator therapy, by improving small airway obstruction, has been shown to reduce lung hyperinflation in stable disease and therefore may partially ‘reset’ the threshold at which exacerbation occurs . This provides a mechanism by which bronchodilator therapy may reduce exacerbation frequency even in the absence of any anti-inflammatory effects.
Triggers of COPD exacerbations
Most exacerbations appear to be associated with infective triggers, either bacterial or viral, although ‘non-infective’ triggers such as air pollution may also be important. Most of the 40% of acute exacerbations linked to viral infections involve rhinovirus (58%) although other respiratory viruses implicated include human respiratory syncytial virus, coronavirus, influenza virus, parainfluenza virus and adenoviruses [24, 25]. These viral associated exacerbations exhibit greater systemic or airway inflammation, greater symptom burden and longer recovery times [9, 11, 24].
Susceptibility to viral infection, COPD and the frequent exacerbator phenotype
An increased viral susceptibility has been proposed as a cause of more frequent exacerbations in COPD . Cigarette smoking does appear to increase viral susceptibility [26, 27] but whether or not the overall COPD population itself is more susceptible is disputed. Data from an analysis in the London COPD cohort showed that upper airway colds are more likely to occur in patients with a history of frequent exacerbations, suggesting that these patients have enhanced susceptibility to viral infection . There is also evidence that patients in whom a respiratory virus is detected have a higher exacerbation frequency . Monto et al., in an early study, demonstrated that patients with chronic bronchitis had a greater incidence of serologically proven rhinovirus infection  although this work involved small study groups, lacked spirometric measurements and preceded molecular viral detection techniques. Further support for the viral susceptibility theory came from Wald et al., who studied a rhinovirus outbreak within a care facility and found that patients with COPD were over-represented in the infected group . However, contradicting these findings is the work by Greenberg et al. who, again using small groups, found no excess incidence of viral respiratory infections in patients with COPD versus controls, though frequent exacerbators were not specifically studied in this paper . Well-powered studies are required to firmly establish whether the frequent exacerbator phenotype in particular experiences a greater proportion of viral exacerbations than infrequent exacerbators.
There are a number of proposed mechanisms of increased viral susceptibility including modulation of intracellular adhesion molecule 1 (ICAM-1). Most rhinovirus serotypes (major group) attach to respiratory epithelial cells by binding to ICAM-1  and through this transmembrane protein modulate the recruitment and activation of inflammatory cells. Increased ICAM-1 increases susceptibility to infection and ICAM-1 appears to be upregulated in the bronchial mucosa of patients with chronic bronchitis .
While viruses are separately thought to be a major cause of COPD exacerbations, 25% of exacerbations involve coinfection with viruses and bacteria. These exacerbations exhibit greater functional impairment, with longer hospitalization [17, 24], and are associated with higher airway bacterial load and greater airway inflammation . Recent work using rhinovirus inoculation as an experimental model of COPD exacerbation shows a consistent increase in bacterial numbers following the initial viral infection  and this supports the hypothesis that bacterial exacerbation may be precipitated by viruses.
Bacterial infection and COPD exacerbations
Common bacteria associated with many COPD exacerbations include Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis and their presence at exacerbation is accompanied by greater levels of systemic inflammation . Besides the introduction of these bacterial species or an increase in their bacterial load, it has also become clear that even a change in bacterial strain can trigger exacerbations .
In the absence of active infection, the identification, by traditional culture methods, of bacteria in the lower airway has been termed ‘colonization’. This term gives the illusion that such bacterial presence is benign, yet it is actually associated with greater inflammation, poorer lung function , chronic bronchitis [39–41], and more frequent exacerbations .
With the development of new molecular techniques it is clear that even the healthy lower airway is not sterile  and so we must re-evaluate the meaning of previously identified colonization and the mechanism by which it influences exacerbation frequency. Colonization may actually represent a disturbance of the normal microbiome, with subsequent overgrowth of a particular species, which drives the inflammatory process in COPD. Alternatively, it may be the inflammatory process that impairs normal regulation of bacterial presence making colonization a marker and potential amplifier of underlying inflammatory damage rather than its primary driver. Longitudinal studies are required to establish how the airway microbiome changes at exacerbation and how this relates to colonization and the frequent exacerbator phenotype.
Genetic predisposition of exacerbation susceptibility
The existence of a seemingly distinct exacerbator phenotype raises the question of whether there is a genetic predisposition towards frequent exacerbations in COPD. This question has not yet been clearly answered, although some distinct gene polymorphisms appear to be linked to exacerbation frequency. Takabatake and colleagues  reported that a single nucleotide polymorphism in the CCL-1 gene, encoding a leucocyte chemotactic factor, was predictive of the frequency and severity of exacerbations, while polymorphic variation in surfactant protein B has also been linked with exacerbation frequency . Mannose-binding lectin (MBL) deficiency is associated with susceptibility to respiratory infections; a higher frequency of MBL gene polymorphisms was noted in patients with COPD who experience frequent exacerbations  and may also predict hospital admission for COPD . In addition to these, an increased exacerbation frequency was found in patients whose sputum contained markers suggestive of increased microsatellite DNA instability, suggesting that those patients with a higher DNA mutation rate experience more frequent exacerbations . Further research is needed to explore this further.
Adherence to medications
It has been suggested that half of all patients do not take their medications adequately , and these patients may be missing the benefit of therapies that would otherwise reduce exacerbation frequency. Treatments for COPD are varied, including rehabilitation, smoking cessation, oxygen therapy as well as medications administered via different routes and devices, and adherence is therefore complex and difficult to measure; it may differ between therapies and even in individual patients . However, patients who are non-adherent to COPD therapy experience more frequent hospitalizations  and have greater healthcare utilization than those who adhere , and therefore non-adherence may be another etiological factor in the development of the frequent exacerbator phenotype. Factors linked to non-adherence include age, complexity of treatment regimen, nature of inhaled delivery device, patient education, and importantly depression , which is itself common in COPD and linked to exacerbation frequency as described below.