Guidelines indicate that pharmacological treatment of chronic obstructive pulmonary disease (COPD) is based on bronchodilators with the addition of inhaled corticosteroids (ICS) in frequent exacerbators [1]. Although this recommendation is in accordance with the best evidence available, it does not take into account that more than half of the exacerbations are infective and infections are not adequately prevented or treated by ICS. In fact, no pharmacological treatment has been able to suppress all exacerbations and, on the contrary, the long-term use of ICS has been associated with higher airway bacterial load in stable COPD [2], and increased risk of pneumonia [3] and tuberculosis [4].
Clinicians have always been aware that not all exacerbations of COPD are alike. Some must be treated with anti-infective agents while others respond better to systemic corticosteroids. Markers such as sputum colour and C-reactive protein levels are useful to differentiate exacerbations and guide therapy [5, 6]. Recent studies have further characterised these different types of exacerbations, and identified bacterial, viral, inflammatory (eosinophilic) and pauci-inflammatory phenotypes of exacerbations, with the majority (59%) of exacerbations being bacterial or viral. Interestingly, the phenotype of the exacerbation remains constant in a given patient [7]. However, despite these different phenotypes of exacerbations, no studies have attempted to investigate whether different treatments would be more effective in preventing one type of exacerbation or another.
Both the high bacterial load in the airways in stable COPD and an increased frequency of exacerbations have been associated with a more rapid decline in lung function and an impairment in quality of life [8–10]. More recently, the presence of bronchiectasis in patients with COPD [11] and the persistent colonisation by unusual microorganisms [12] have also …