A cough is an important reflex in health that protects the lungs from aspiration and clears secretions. However, coughing can also be disruptive to the patient and is the most common reason why patients consult their doctor. A wide number of pulmonary and extrapulmonary conditions can lead to a persistent cough. Cough can be broadly divided into three categories: acute, subacute and chronic cough.1 Acute cough is defined as cough with a duration of < 3 weeks and is the most common type of cough, usually caused by viral upper respiratory tract infections. Subacute cough is defined as cough duration of between 3 and 8 weeks and is most commonly a sequel of a viral upper respiratory tract infection. Chronic cough, defined as a cough with a duration of > 8 weeks, is a common complaint associated with significant impairment in health-related quality of life and patients are often referred to respiratory outpatient clinics. Numerous international cough guidelines have been published to guide the management of patients with chronic cough;2–4 however, a significant number of patients continue to suffer from a cough despite extensive investigations and treatment trials.5 This review will focus on the management of cough, recent developments and also highlight areas of controversy.
The neural basis for cough
A number of receptors located in the airways and lungs have been identified as potential cough receptors. They are thought to be most concentrated in the laryngopharyngeal area. Cough receptors and their afferent nerves can be simply divided into chemoreceptors and mechanoreceptors. Mechanoreceptors, such as the rapidly adapting receptor or ‘cough receptor’, are activated by triggers such as secretions and foreign bodies. Chemosensitive receptors protect against irritants and noxious triggers and examples include transient receptor potential vanilloids and ankyrin type 1 receptors, TRPV1 and TRPA1. They are activated by triggers such as capsaicin, acid, heat, pungent vapours and cigarette smoke. Cough receptors are located on the nerve endings innervating airway epithelium and the nerve fibres are carried in the vagus nerve to the nucleus tractus solitarii in the brain stem where the cough reflex is coordinated (Figure 1). The cortex is also important as it contains areas that are involved in stimulus localization, perception and intensity and voluntary cough suppression.6 A key finding in most types of cough is heightened cough reflex sensitivity and, although recent research has improved the understanding of cough receptors and their afferent nerve pathways, the mechanism of the sensitization of the cough reflex remains elusive. A better understanding will lead to the development of novel antitussive medications.
The sensitivity of the cough reflex can be assessed by inducing cough with tussive agents, such as capsaicin or citric acid.7 They are usually nebulized in increasing doses until five or more coughs are induced; however, there are two significant limitations with this methodology. First, there is wide overlap of cough reflex sensitivity between healthy subjects and patients with cough;7 hence, this test has little diagnostic value to identify pathological cough reflex hypersensitivity. Second, the assessment of cough reflex hypersensitivity is poorly predictive of a positive response to trials of therapy.8 It is possible that these limitations are related to the protocols used rather than the concept of assessing cough reflex sensitivity. A recent study has demonstrated it is possible to discriminate healthy subjects from those with cough by administering higher doses of tussive agents than currently used.9 Future research should refine the methodology of cough challenge tests, so that they are clinically useful. The distinction between cough reflex hypersensitivity and other causes of cough, such as airway hyper-responsiveness, inflammation and infection, may be useful to guide therapy and should be explored further.
Acute cough – viral upper respiratory tract infection
Acute cough is most often caused by a viral illness. Viral infection of the upper respiratory tract leads to a transient increase in cough reflex sensitivity, and the presence of associated symptoms such as a sore throat, rhinorrhoea and fever is highly suggestive of an upper respiratory tract infection. Most patients can be managed with supportive care, rather than pharmacological therapy, but in some patients the cough is intense, disrupts work or sleep and is associated with adverse symptoms such as chest pain and vomiting.
A significant number of patients use over-the-counter antitussive therapy. In the USA, sales of antitussive drugs exceed 6 billion US dollars each year, despite little evidence for their efficacy and the recommendation against their use from The American College of Chest Physicians.2 The author’s recommendation is simple cough linctus as the first-line therapy. The syrup has a demulcent effect and the sugary taste encourages hypersalivation, which may also alleviate symptoms.10 Simple linctus is safe and may be the only intervention necessary for many patients.
A wide range of medications is added to cough suppressants, such as dextromethorphan, antihistamines and decongestants, often in combination. The author’s recommendation for patients who have not responded to supportive care and simple linctus is to give cough suppressants containing dextromethorphan, as the evidence for efficacy is better than the alternatives. Dextromethorphan is a centrally acting N-methyl-d-aspartate receptor antagonist and was approved by the Food and Drug Administration, MD, USA, in 1968 for cough suppression based largely on the findings of studies of small numbers of hospitalized patients with tuberculosis. A subsequent study reported a 17% reduction in cough frequency with dextromethorphan in cases of acute cough;11 however, the clinical relevance of this improvement is questionable. Dextromethorphan has been withdrawn for children under the age of 6 years in the UK and USA owing to concerns regarding the lack of evidence to support efficacy and side-effects. There are significant cultural variations in the use of antitussive drugs; for example, levodropropizine, an antitussive with an unknown mechanism of action, is widely used in Italy, but not the UK. A large number of medications are used in cough remedies and there is a need for better understanding of their mechanism of action and efficacy. Some antitussive medications are marketed for productive cough and examples are mucolytics such as carbocisteine and expectorants such as guaifenesin. As with other antitussives, there is very limited evidence for their efficacy and further studies are required to address this.
Subacute cough – post-infectious cough
Subacute cough is most commonly a sequel of a viral upper respiratory tract infection. Whooping cough (pertussis) due to Bordetella pertussis infection is also an important cause in adult patients.12 Most patients with post-viral subacute cough are readily identified from a history of a recent upper respiratory tract infection and do not require investigation. The differential diagnosis includes asthma and rhinitis. The treatment for most patients is supportive care. If further management is necessary, the options are guided by expert opinion as with acute cough, rather than being evidence based. The author’s suggestion is to use a similar strategy to acute cough: simple linctus followed by cough suppressant medications containing dextromethorphan.
B. pertussis infection in adults is an important and under-recognized cause of subacute cough. In the UK, pertussis has a prevalence of 25% of subacute cough when confirmed with objective laboratory testing. The presence of paroxysmal cough, whoop and post-tussive emesis is suggestive of subacute cough and, in adults, atypical presentations without associated symptoms are common. Treatment is most effective in reducing the severity of symptoms if offered early in the disease course, before the onset of paroxysms of coughing. Unfortunately, this is the time when pertussis is most difficult to diagnose. The suspicion of pertussis infection may be increased during known outbreaks and, when there is a suspicion of pertussis, macrolide antibiotics are recommended if the patient presents within 3 weeks of onset. There are no effective treatments that reduce pertussis-induced cough severity. The current evidence does not support the use of corticosteroids, beta-agonists, montelukast or antihistamines. The diagnosis of pertussis using laboratory tests is challenging owing to the limitations of the methodology available. Diagnostic tests include culture (delayed result), direct fluorescent antibody testing (poor sensitivity and specificity), serology (retrospective diagnosis owing to the requirement of a second blood test) and polymerase chain reaction. For most patients, the initiation of antibiotics will largely be based on symptoms that suggest pertussis and if there is a known epidemic in the local community.
Patients with chronic cough are usually referred to respiratory outpatient clinics, but can also present to a number of other specialties, such as ENT (otolaryngology), gastroenterology and allergy. Chronic cough is associated with significant physical and psychological morbidity.13 Adverse symptoms associated with cough include chest pain, syncope, urine incontinence and depression. Chronic cough can be disruptive in the workplace and social environment. A wide range of pulmonary and extrapulmonary conditions can cause chronic cough. Pulmonary conditions such as chronic obstructive pulmonary disease (COPD), tuberculosis, lung cancer, pulmonary fibrosis and sarcoidosis are easily identified with chest radiology and lung function tests. This review article will focus on patients with an unexplained chronic cough.
Investigating patients – the anatomic diagnostic protocol
The anatomic diagnostic protocol (ADP) was described over 30 years ago to guide the investigation and management of patients suffering from chronic cough.14 The ADP is based on the finding that chronic cough is commonly associated with asthma, gastro-oesophageal reflux disease (GORD) and rhinitis.15 The diagnosis is usually established following trials of therapy directed at these three conditions. Early reports suggested a very high success rate (95%) using the ADP to diagnose and treat cough,16 but recent studies have reported less favourable outcomes, with a prevalence of unexplained cough as high as 40%.5 However, the ADP remains central to the approach recommended by cough guidelines and is discussed below. The limitations and the need to develop new approaches to manage patients are also discussed.
The history taken should evaluate factors surrounding the onset of cough such as an upper respiratory tract infection, triggers of cough, adverse quality of life (particularly urine incontinence), presence of sputum production, breathlessness and wheeze, and this is summarized in Table 1. The triggers of cough and the activation of neural mechanisms suggest that coughing has a neuropathic basis. Triggers can be divided into known tussigens, for example smoke and scents (hypertussia), and non-tussigens, such as talking (allotussia) (see Table 1).17 Up to 75% of patients experience a tickle or other sensations in the throat (laryngeal paraesthesia), which is suggestive of heightened cough reflex sensitivity. The patient’s history should record the presence of GORD, rhinosinusitis, snoring and numbness (peripheral neuropathy)18 and the examination should focus on the upper airways, respiratory and cardiovascular systems. The physician should be vigilant when examining for the presence of enlarged tonsils, external ear wax (e.g. Arnold’s reflex) and premature ventricular contractions. Most patients require only chest radiography and spirometry as initial investigations.
The initial management of the patient should focus on the most likely aetiology. If an underlying cause is not apparent, then the patient should undergo a therapeutic trial for eosinophilic cough (also known as cough-variant asthma or eosinophilic bronchitis), as this responds well and consistently to specific therapy. The management of conditions commonly associated with a chronic cough is described in Table 2.
Eosinophilic cough (cough-variant asthma/eosinophilic bronchitis)
Eosinophilic cough accounts for approximately 25% of all cases of chronic cough. A cough is usually the only symptom and is associated with eosinophilic airway inflammation. In contrast with classic asthma, the cough is associated with heightened cough reflex sensitivity and the diagnosis can be established in the laboratory with tests for bronchial hyper-responsiveness and airway inflammation (exhaled nitric oxide or induced sputum differential cell count). The presence of a positive bronchial responsiveness test in isolation is not indicative of cough-variant asthma: evidence of airway inflammation is also necessary. When tests for airway inflammation are not available, a trial of corticosteroid therapy is recommended.4 In an otherwise well patient, a low-dose, short-duration trial of oral corticosteroids, such as 25 mg prednisone once daily for 7 days, is suggested. This is less than the 2 weeks duration recommended by cough guidelines but, in the author’s experience, a longer duration is not necessary as most patients respond within 7 days. A short trial of oral corticosteroids may be preferable to inhaled corticosteroids for many patients because of multiple limitations when using the latter. Trials of inhaled corticosteroids frequently fail because of poor technique, inadequate dose, inadequate duration, inhaler-induced cough and poor compliance. Caution is advised in using oral corticosteroids in the elderly or obese and in diabetic patients and patients with impaired glucose intolerance. Cough-variant asthma is usually treated with inhaled corticosteroids and a suggested therapeutic option is a 400 μg dose of inhaled beclomethasone equivalent twice daily via a spacer device. An alternative to inhaled corticosteroids is leukotriene antagonists, for which there is some evidence of efficacy.19
Eosinophilic bronchitis is another important cause of chronic cough, reported to have a prevalence of between 5% and 15%.20 In contrast to cough-variant asthma, there is an absence of airway hyper-responsiveness. The cough is thought to result from mast cells infiltrating the airway epithelium as opposed to the airway smooth muscle infiltration found in classic asthma.21 Similar to cough-variant asthma, patients with eosinophilic bronchitis have heightened cough sensitivity. Eosinophilic bronchitis can be identified by demonstrating the presence of eosinophilic airway inflammation (exhaled nitric oxide or induced sputum cell analysis) and the absence of bronchial responsiveness. When tests for bronchial responsiveness or airway inflammation are not available, a trial of corticosteroid therapy is warranted. Distinction between cough-variant asthma and eosinophilic bronchitis will not be possible with trials of therapy, but this may have little impact on the management of the patient. The long-term treatment of eosinophilic bronchitis, as with cough-variant asthma, is to prescribe inhaled corticosteroids.22
Gastro-oesophageal reflux has been reported to cause up to 40% of cases of chronic cough. It is thought to cause cough by an oesophageal–bronchial reflex, rather than microaspiration into the lungs.23 The diagnosis of gastro-oesophageal reflux cough is challenging because the presence or absence of symptoms does not always lead to confirmation or exclusion of the diagnosis. A 24-hour oesophageal pH test is a poor predictor of the diagnosis, when judged by the response to therapy.24 ENT specialists often diagnose reflux-associated cough, also known as laryngopharygeal reflux (LPR), following laryngoscopic appearances of vocal cord oedema and erythema; however, this is neither specific for reflux nor predictive of a response to therapy. The diagnosis of reflux cough is frequently based on the response to a trial of treatment. In the author’s opinion, a good response to a trial of treatment should be followed up with a trial of discontinuation of treatment once the cough has settled. This is because placebo responses in trials of cough can be significant.25 The diagnosis of non-acid gastro-oesophageal reflux-associated cough is even more challenging and a 24-hour oesophageal impedance is the preferred test for many clinicians. However, a recent study reported that the volume of non-acid reflux is not increased in patients with cough compared with healthy subjects.26 It may be possible for physiological levels of reflux to cause cough in sensitized patients. A temporal association between cough and non-acid reflux is needed to support this hypothesis, but a recent study reported that cough was as likely to precede episodes of reflux as to follow it.23 In clinical practice, it is difficult to demonstrate a temporal relationship because fully integrated reflux and cough monitoring tools are not widely available outside research centres. Identifying cough with oesophageal manometry is flawed as this is an insensitive method, missing > 30% of coughs.23 The author recommends that oesophageal impedance studies should be limited to patients with troublesome therapy-resistant heartburn rather than for an isolated cough. Furthermore, oesophageal impedance study analysis is not fully automated and, therefore, is relatively time-consuming. The role of assessing oesophageal dysmotility with manometry is unclear. Disorders of oesophageal motility are a frequent finding in patients with chronic cough, but the diagnostic value of such disorders is questionable.
Acid gastro-oesophageal reflux-associated cough is treated with proton pump inhibitor (PPI) therapy and an example is 40 mg omeprazole twice daily for 2 months. It is important to note that two well-conducted randomized controlled trials25,27 in patients with chronic cough who take high-dose PPI therapy were convincingly negative. Furthermore, a meta-analysis of trials of PPI in patients with LPR did not find an overall benefit compared with the placebo.28 These studies suggest that acid gastro-oesophageal reflux is not the cause of chronic cough, at least in the specialist clinic setting. The role of PPI therapy in cases of cough is uncertain at present and further guidance on this issue is awaited in forthcoming revisions of cough management guidelines. In the author’s opinion, a single trial of PPI therapy should be recommended for patients with symptoms of gastro-oesophageal reflux until further guidance is available. This is because of the possibility that some patients may be successfully treated with PPIs in primary care and, therefore, do not participate in clinical trials. There are no effective pharmacological treatments available for non-acid gastro-oesophageal reflux-associated cough. Metoclopramide and domperidone are frequently used to treat oesophageal dysmotility, but there are no randomized controlled trials to support their efficacy. Surgery such as Nissen fundoplication is sometimes recommended and a small number of case series have suggested that this procedure is effective; however, the long-term outcome of patients does not appear to differ significantly from those undergoing a conservative approach.29 Oesophageal surgery is associated with noteworthy morbidity and side-effects and, in the author’s opinion, surgery should be reserved for patients with severe, refractory and symptomatic gastro-oesophageal reflux, rather than solely for the symptoms of cough.
The upper airways – rhinitis and post-nasal drip
Upper airway symptoms such as nasal congestion, rhinorrhoea and post-nasal drip are frequently present in patients with a cough. In most people, they are not as troublesome as the cough and are usually identified following directed questioning. Activation of laryngeal mechanoreceptors and extension of inflammation into the laryngopharnyx are potential mechanisms for cough. The role of rhinitis in cough remains an area of controversy30 and there is paucity of evidence for a causal association. Furthermore, ENT specialists are more likely to attribute the cough to LPR rather than rhinitis or post-nasal drip.
The need for upper airway investigations should be based on the severity of rhinitis rather than cough. This is because they are not predictive of a reduction in cough severity following therapy. When appropriate, investigations should include allergy skin-prick tests (or immunoglobulin E blood tests) to identify the presence of atopy and a referral to an ENT specialist for laryngoscopy and a computed tomography scan of sinuses. The latter should be reserved for selected cases that do not respond to initial therapy. The evaluation of the upper airways in patients without upper airways symptoms is of limited value. The treatment of rhinitis-associated cough is carried out with nasal corticosteroids, as either a spray or drops, and antihistamines. The first-generation sedating antihistamine, dexbrompheniramine, has been the most studied but is not widely available.1 It is known to have antitussive properties and it is perhaps this, rather than an antihistamine effect, that may be the mechanism of action.31
Other associations with cough
Angiotensin-converting enzyme (ACE) medication can cause or aggravate a cough and should always be discontinued if the cough affects quality of life. ACE antagonists are an alternative as they do not cause cough. Cough may be the sole presenting feature of obstructive sleep apnoea32 and this association should be considered for patients who have a raised body mass index, who snore and cough at night and when other more common causes have been excluded. An overnight oximetry test should be the first line of investigation and may be all that is necessary to confirm the diagnosis; treatment with continuous positive airway pressure has been reported to be effective.33 Chronic tonsillar enlargement has been associated with unexplained chronic cough in a small study34 in which tonsillectomy led to a symptomatic improvement and reduction in cough reflex sensitivity. Premature ventricular contractions are a rare cause of cough but this should be considered in patients with symptoms of palpitations and irregular pulse. Twenty-four-hour echocardiography should be the first line of investigation in such cases and pharmacological treatment or cardiac ablation therapy has been reported to reduce the severity of the cough.35 B. pertussis infection is emerging as an important cause of chronic cough in adults and has been discussed above. Other associations with cough include snoring and ear canal obstruction/wax (Arnold’s reflex).
Limitations of the anatomic diagnostic protocol
The success rate of the ADP is lower than previously reported5 and, in the author’s opinion, although the ADP has led to much deserved recognition of patients with cough, it has also led to some limitations. Many patients undergo too many, sometimes unnecessary, investigations and some of which are invasive. Multiple hospital clinic visits are a significant inconvenience for patients. The ADP has inadvertently led to an unbalanced focus on the symptoms associated with cough, rather than the cough itself. There has been an increase in research investigating the mechanisms of cough reflex hypersensitivity and this reflects the importance of cough receptors, afferent nerves and the brain in regulating cough. This approach may identify novel targets for therapeutic intervention.
The management of chronic refractory cough
The therapeutic options for patients with a persistent cough following thorough evaluation are very limited. A suggested approach is outlined below based on the available evidence.
Cough suppression therapy. A self-management approach has been recently described by speech and language therapists and respiratory physiotherapists.36 The programme consists of 2–4 sessions and encourages patients to suppress their cough voluntarily. The components of the programme include education, laryngeal hydration and hygiene, cough control and psychoeducational counselling. The education component alerts patients to the negative effects of repeated coughing and reassurance that cough suppression is not harmful. The laryngeal component identifies factors that may cause laryngeal irritation, such as mouth breathing and smoky atmospheres, and encourages patients to hydrate their larynx. Patients are taught to control their cough by either suppressing it or replacing it with a forced swallow or distraction techniques such as sipping water or sucking a sweet. Paradoxical vocal fold movement is frequently found in patients with chronic cough; therefore, this is identified and addressed, as it can exacerbate cough. In addition, a number of breathing control and retraining techniques can be used to address this. Interventions to reduce laryngeal muscle tone are also encouraged. Psychoeducational counselling aims to modify patient behaviour, reduce stress and anxiety and encourages patients to take control of their cough. Studies evaluating cough suppression therapy have reported very encouraging results. A randomized controlled trial carried out by speech therapists and an observational study led by physiotherapists demonstrated a reduction in cough symptoms and an improvement in the quality of life, respectively.37,38 There is much overlap between speech and physiotherapy techniques and the differences between them may be an increased focus on voice disorders and dysfunctional breathing, respectively.
Gabapentin. Cough is considered to be a neuropathic condition, similar to neuropathic pain.17 This is because there are similarities in the characteristics of symptoms, nerve receptors and their pathways. Gabapentin is thought to act centrally by inhibiting calcium channels and possibly N-methyl-d-aspartate receptors. A recent randomized controlled trial in patients with chronic refractory cough demonstrated a reduction in cough frequency and improvement in health-related quality of life.8 Patients were given a dose up to 600 mg, three times a day. The treatment duration was 2 months but the cough returned following discontinuation of therapy. The long-term effects of gabapentin therapy in patients with cough are not known. Limitations of gabapentin therapy include the occurrence of central nervous system side-effects, such as drowsiness, and the difficulty in predicting which patients will respond. An alternative to gabapentin therapy is low-dose amitriptyline as uncontrolled studies have reported an antitussive effect.39 The starting dose is 10 mg at night, but this can be increased to 25 mg or 50 mg. Side-effects include drowsiness and a dry mouth. A randomized controlled trial of another neuromodulator, pregabalin, is currently under way in the Department of Respiratory and Sleep Medicine, Hunter Medical Research Institute, John Hunter Hospital, Newcastle, New South Wales, Australia.
Morphine. A randomized controlled trial of 5 mg morphine sulphate, twice daily, reported an improvement in health-related quality of life.40 Patients tolerated morphine well in this study but morphine has not been widely used in patients with refractory cough, possibly owing to concerns about dependence and opiate side-effects.
Cough hypersensitivity syndrome – the new paradigm
The author’s opinion is that chronic cough should be viewed as a condition and not merely a symptom. Cough reflex hypersensitivity is a key mechanism and the term cough hypersensitivity syndrome (CHS) has been proposed by a European Respiratory Society Taskforce.41 The proposed definition for CHS is ‘a clinical syndrome characterised by troublesome coughing, often triggered by low levels of thermal, mechanical or chemical exposure’. CHS may mimic or coexist with other pulmonary or extrapulmonary diseases and it is associated with cough reflex hypersensitivity and is thought to be distinct from methocholine bronchial hyper-responsiveness. As with other syndromes, such as asthma and COPD, it is likely to be subdivided into phenotypes. CHS is more prevalent in middle-aged women and patients often present with a dry tickly cough. Cough reflex hypersensitivity can be assessed via capsaicin or citric acid cough challenge tests. The aim of therapy is to reduce the cough hypersensitivity and this may be possible in some patients by treating associated conditions, such as eosinophilic airway inflammation. In many patients, however, cough hypersensitivity may be the only abnormality. Currently, there are few effective antitussive therapies, but there are many clinical trials of potential antitussives under way. Potential targets include peripheral airway receptors on nerve endings such as P2X3, TRPA1 and selective sodium channel inhibitors.
Chronic cough or, more appropriately, cough hypersensitivity syndrome is a common condition. Cough reflex hypersensitivity is a key mechanism and the management of patients currently focuses on reducing aggravators that may exacerbate a cough. There is a pressing need for effective antitussive therapy and a number of potential cough receptors have been identified as therapeutic targets. Future research should focus on identifying the mechanisms of sensitization of the cough reflex.
Morice AH, McGarvey L, Pavord I. Recommendations for the management of cough in adults. Thorax 2006; 61(Suppl. 1):i1–24.
Birring SS. Controversies in the evaluation and management of chronic cough. Am J Respir Crit Care Med 2011; 183:708–15.
Irwin RS, Madison JM. The diagnosis and treatment of cough. N Engl J Med 2000; 343:1715–21.