Among the most common causes of chronic cough are asthma (25%) and nonasthmatic eosinophilic bronchitis (10%). In asthma, cough may present as an isolated symptom, in which case it is known as cough variant asthma. Variable airflow obstruction and airway hyper-responsiveness are cardinal features of asthma, which are absent in nonasthmatic eosinophilic bronchitis. The presence of eosinophilic airway inflammation is a common feature of asthma and is a diagnostic criterion for nonasthmatic eosinophilic bronchitis. At a cellular level, mast cell infiltration into the airway smooth muscle bundle, narrowing of the airway wall, and increased interleukin-13 expression are features of asthma and not nonasthmatic eosinophilic bronchitis. In most cases, the trigger that causes the cough is uncertain, but occasionally occupational exposure to a sensitizer is identified, and avoidance is recommended. In both conditions, there is improvement following treatment with inhaled corticosteroids, which is associated with the presence of an airway eosinophilia and increased exhaled nitric oxide. Generally, response to therapy in both conditions is very good, and the limited long-term data available suggest that both usually have a benign course, although in some cases fixed airflow obstruction may occur.
Chronic cough is defined as a cough lasting for more than 8 weeks with no clinical or radiological evidence of lung disease; it usually results in referral for specialist assessment. An anatomic diagnostic protocol has been constructed in several studies, and this protocol helps identify the causal process in most of these patients. Asthma accounts for about one in four cases of cough. Sputum induction was one the first validated noninvasive techniques to assess airway inflammation and led to the identification of nonasthmatic eosinophilic bronchitis. This condition manifests like asthma with sputum eosinophilia and accounts for 10% of cases of chronic cough, but unlike asthma, there is no airway dysfunction. This article summarizes the current understanding of the similarities and differences between asthma and nonasthmatic eosinophilic bronchitis in terms of the immunopathogenesis, natural history and clinical management. The main features and differences between the asthma, cough variant asthma, and nonasthmatic eosinophilic bronchitis are as shown in Table 1 .
| Eosinophilic Bronchitis | Classical Asthma | Cough Variant Asthma | |
|---|---|---|---|
| Symptoms | Cough, often upper airway symptoms | Breathlessness, cough, wheeze | Isolated cough |
| Presence of atopy | ↔ | ↑ | ↑ |
| Airway hyper-responsiveness | – | + | + |
| Response to bronchodilator | – | + | + |
| Cough reflex hypersensitivity | ↑ | ↔ | ↑ or ↔ |
| Sputum eosinophil count | ↑ | ↑ or ↔ | ↑ or ↔ |
| Mast cells within airway smooth muscle bundles | – | + | + |
| Response to corticosteroids | Good | Good in presence of sputum eosinophilia | Good in presence of sputum eosinophilia |
Definition, diagnosis, and prevalence
Asthma is the leading cause for chronic cough, accounting for 24% to 29% cases, and these percentages were fairly similar in multiple prospective studies of adult nonsmokers. Classic asthma presents most commonly with cough, wheeze, and dyspnea and chest tightness. Cough variant asthma, as the title suggests, presents with cough in isolation. Asthmatic patients may have normal physical examination and spirometry; hence when a cause for cough is being sought, bronchial challenge testing may reveal the presence of reactive airways disease. Antiasthma treatment in the form of bronchodilators, inhaled corticosteroids, or a short course of oral prednisolone (usually 0.5 mg/kg for 2 weeks), may resolve the cough, which confirms the diagnosis.
Nonasthmatic eosinophilic bronchitis is defined as a chronic cough in patients with no symptoms or objective evidence of variable airflow obstruction, normal airway hyper-responsiveness (provocative concentration of methacholine producing a 20% decrease in forced expiratory volume in 1 minute (FEV 1 ), PC 20 greater than 16 mg/mL), and sputum eosinophilia. An accepted upper cut-off level of greater than 3% nonsquamous sputum eosinophils is used as indicative of eosinophilic bronchitis, as this is outside the 90th percentile for normal patients (1.1%). This level of sputum eosinophilia has been associated with a corticosteroid response in asthma and chronic obstructive pulmonary disease (COPD).
Importantly, another cause of chronic cough has been described, particularly in Japanese populations, that is characterized by the presence of eosinophilic airway inflammation and atopy, without airway hyper-responsiveness known as atopic cough. This condition is very similar to nonasthmatic eosinophilic bronchitis, except that the latter includes patients with and without atopy. It is therefore likely that atopic cough represents a subgroup of nonasthmatic eosinophilic bronchitis rather than a distinct condition, although further comparative studies are required.
The information gained by measuring airway inflammation only recently has been more widely recognized. Herein lies the problem with establishing the true prevalence of nonasthmatic eosinophilic bronchitis, as most reports of the causes of chronic cough did not include these investigations. In the few studies where airway inflammation was measured, however, it accounted for 10 to 30% of cases of chronic cough. In a 2-year prospective study of chronic cough at a specialist center, 91 patients were identified among an initial 856 referrals. A sputum induction was performed on carefully selected patients, those in whom the diagnosis was unclear despite examination and initial testing. This led to the identification of 12 (13.2%) patients who had nonasthmatic eosinophilic bronchitis, representing about 30% of patients who underwent sputum induction.
Exhaled nitric oxide as a surrogate for sputum eosinophilia
Sputum induction and analysis require technical support and training, which have limited their widespread application. A simpler, cheaper near-patient biomarker of eosinophilic inflammation is therefore attractive. Exhaled nitric oxide (eNO) levels are increased in asthma, cough variant asthma, and nonasthmatic eosinophilic bronchitis, and expression of inducible nitric oxide synthase is up-regulated in the epithelium of asthmatics. Both of these features respond to corticosteroid therapy. The correlation between eNO and sputum eosinophilia is good in corticosteroid-naïve patients, but this relationship is weaker once patients are treated. In asthma following treatment with anti-interleukin (IL)-5, the sputum eosinophil count is markedly attenuated without affecting eNO, suggesting that this relationship represents an epiphenomenon rather than a causal association. The application of eNO to guide corticosteroid therapy in asthma has been disappointing compared with success with sputum eosinophil counts and therefore questions its role in monitoring patients and titrating their therapy. This, however, does not detract from its potential as a biomarker for eosinophilic inflammation during the workup of patients as part of the diagnostic investigations. Indeed, Hayn and colleagues reported that eNO had good positive and negative predictive values for response to corticosteroids in chronic cough with a sensitivity and specificity above 85%. Similarly, eNO has a strong negative predictive value for a sputum eosinophilia in nonasthmatic eosinophilic bronchitis and for cough variant asthma. It therefore probably has clinical utility as part of the initial evaluation for chronic cough before the use of corticosteroid therapy for the exclusion of cough variant asthma and nonasthmatic eosinophilic bronchitis. Therefore, although sputum induction and processing would be desirable in the routine assessment of chronic cough, eNO does provide valuable information in cases where analysis of sputum is not possible.
Exhaled nitric oxide as a surrogate for sputum eosinophilia
Sputum induction and analysis require technical support and training, which have limited their widespread application. A simpler, cheaper near-patient biomarker of eosinophilic inflammation is therefore attractive. Exhaled nitric oxide (eNO) levels are increased in asthma, cough variant asthma, and nonasthmatic eosinophilic bronchitis, and expression of inducible nitric oxide synthase is up-regulated in the epithelium of asthmatics. Both of these features respond to corticosteroid therapy. The correlation between eNO and sputum eosinophilia is good in corticosteroid-naïve patients, but this relationship is weaker once patients are treated. In asthma following treatment with anti-interleukin (IL)-5, the sputum eosinophil count is markedly attenuated without affecting eNO, suggesting that this relationship represents an epiphenomenon rather than a causal association. The application of eNO to guide corticosteroid therapy in asthma has been disappointing compared with success with sputum eosinophil counts and therefore questions its role in monitoring patients and titrating their therapy. This, however, does not detract from its potential as a biomarker for eosinophilic inflammation during the workup of patients as part of the diagnostic investigations. Indeed, Hayn and colleagues reported that eNO had good positive and negative predictive values for response to corticosteroids in chronic cough with a sensitivity and specificity above 85%. Similarly, eNO has a strong negative predictive value for a sputum eosinophilia in nonasthmatic eosinophilic bronchitis and for cough variant asthma. It therefore probably has clinical utility as part of the initial evaluation for chronic cough before the use of corticosteroid therapy for the exclusion of cough variant asthma and nonasthmatic eosinophilic bronchitis. Therefore, although sputum induction and processing would be desirable in the routine assessment of chronic cough, eNO does provide valuable information in cases where analysis of sputum is not possible.
Pathogenesis
The etiology remains unknown, but both asthma and nonasthmatic eosinophilic bronchitis can be associated with exposure to inhaled aeroallergen or an occupational sensitizer. These conditions share many immunopathological features including a similar degree of sputum, bronchoalveolar lavage, and biopsy eosinophilia, and a similar degree of reticular basement membrane thickening. Similarly, there are increased sputum concentrations of cysteinyl–leukotrienes and eosinophilic cationic protein. IL-5 and granulocyte macrophage colony stimulating factor gene expression are increased in bronchoalveolar lavage in both asthma and nonasthmatic eosinophilic bronchitis. The role of vascular endothelial growth factor (VEGF) is unclear, as one study found VEGF levels similar in both asthma and nonasthmatic eosinophilic bronchitis, while another study showed higher sputum VEGF concentration in the asthma group. This is implicated in causing airway narrowing by increasing vascular permeability and subsequently mucosal edema.
In asthma, mast cell numbers in airway smooth muscle were increased, but not in nonasthmatic eosinophilic bronchitis, and the number of airway smooth muscle mast cells was correlated inversely with airway hyperresponsiveness. Histamine and prostaglandinD2 sputum concentrations are increased in non-asthmatic eosinophilic bronchitis. This supports the view that mast cells are localized and activated in more superficial airway epithelium in nonasthmatic eosinophilic bronchitis, whereas in asthma they lie deeper in close association with the airway smooth muscle. IL-13 expression is increased in asthma in bronchial submucosa, sputum, and peripheral blood T-cells.
The airway narrowing that characterizes asthma is not seen in nonasthmatic eosinophilic bronchitis. It is a consequence of airway remodeling, which is active in both conditions but only results in disordered airway geometry in the former. Thus, although there is an overlap in the immunopathogenesis of both conditions, the pathways of airway inflammation are likely to ultimately lead to the differences in clinical expression between asthma and nonasthmatic eosinophilic bronchitis. In summary, airway hyper-responsiveness in asthma is caused in part by the localization of activated mast cells to airway smooth muscle bundle, as evidenced by increased IL-13 expression and airway narrowing, whereas cough is mediated in part by infiltration into the superficial airway by eosinophils and mast cells. These key differences in the immunopathology of asthma and nonasthmatic eosinophilic bronchitis and their clinical consequences are illustrated in Fig. 1 .
