To illustrate reasonable and cost-effective management of cough, one of the most frequent reasons for primary care consultations.
To assist in the secondary care diagnosis and treatment of chronic cough as the solely presenting symptom if chest radiography and lung function tests remain inconclusive.
To emphasise the rational order starting with simple (noninvasive and cost-effective) diagnostic procedures, graduating to complex, invasive and expensive (e.g. computed tomography and bronchoscopy) measures.
Summary Cough is highly prevalent as a cumbersome presenting complaint for many patients. The symptom cough is elicited by a myriad of very different respiratory and nonrespiratory diseases. Symptomatic pharmacological treatment is of very limited efficacy. Presently, no new drug for cough in the pipeline has completed phase II development. Therefore, treating cough requires exact diagnosis for causal treatment. Cough is often the first (but not necessarily an early) symptom of life-threatening diseases, such as lung tumours or recurrent pulmonary embolism. Thus, in 2004, the German Respiratory Society published evidence-based guidelines on the management of cough. Acute and chronic cough were defined, and algorithms were provided for diagnostic workup. In 2010, the guidelines were updated (http://leitlinien.net). An abridged english version is also now available . In this review, evidence-based recommendations (mostly weak evidence or consensus) from these guidelines are adapted into learning points.
Cough is both an important physiological reflex protecting the airways, and a frequent complaint associated with virtually all pulmonary and several extrapulmonary diseases. Cough is also a contributing factor in the spreading of infectious disease, such as tuberculosis. Moreover, (acute) cough due to the common cold is one of the most frequent causes of primary care consultations.
The reflex is characterised by complexity and plasticity, and is triggered by physical and chemical stimuli. Irritant receptors and C-fibre receptors are activated in the airways, pleura, pericardium and oesophagus. The impulse is then transmitted to the brainstem cough generator circuit via the vagus nerves. There is also a connection to the cortex, allowing voluntary control of both eliciting and, to a limited degree, inhibiting cough . Efferent innervations reach the effector muscles (diaphragm, abdominal, intercostals, back, and muscles of the larynx and upper airway). Mucociliary clearance is the primary means of clearing the bronchial system. To a certain degree, cough can compensate for impaired mucociliary clearance (e.g. that caused by the effects of smoking). If mucociliary clearance is overwhelmed by aspiration, an intact cough reflex protects the lungs effectively. However, an impaired cough reflex, e.g. after stroke, results in life-threatening aspiration pneumonia. The clearing competence of the cough reflex depends on several conditions: obstruction of the airways, bronchial collapsibility, lung volumes, respiratory muscle and laryngeal function, and the amount and viscosity of the mucus .
Cough is productive (wet) if the amount of the daily expectoration is ≥30 mL (two tablespoons’ worth). The phlegm can be mucous, serous, purulent or bloody. Bronchial casts can also be produced.
The cough reflex arc consists of five parts: 1) cough receptors; 2) afferent nerves; 3) brainstem cough generator circuit; 4) efferent nerves; and 5) effector organs (muscles).
Hypersensitivity of the cough nociceptors elicits pathological cough. Numerous respiratory and other diseases cause cough nerve hypersensitivity and thus, produce cough. In a considerable number of clinical cases, however, only cough nerve sensitivity is affected without another “specific” cause. Such patients are suffering from idiopathic cough.
Classification and frequent causes of cough
Acute and chronic cough
Cough as a symptom is attributed to distinct diseases and is categorised as either acute (lasting ≤8 weeks) or chronic (lasting >8 weeks). Of course, these limits are arbitrary. Acute cough (due to common cold) usually lasts only 2–3 weeks. Various acute infections, (e.g. Mycoplasma pneumoniae and adenoviruses), however, can elicit cough lasting ≤8 weeks; for example, Bordetella pertussis can cause cough lasting ≤3 months. In otherwise healthy individuals, acute infections of the upper and/or lower airways, the most common cause of cough, are self-limiting. Medical history and physical examination are usually sufficient in the diagnosis, and over-the-counter (OTC; i.e. nonprescription) remedies for the treatment. However, a few special circumstances require immediate, full diagnosis of acute cough (table 3).
In contrast, immediate diagnostic work-up is essential in all patients presenting with chronic (>8 weeks duration) cough; a chest radiograph and lung function test should be performed immediately. This is consistent with the recommendations in all published guidelines on cough [4–7].
If the chest radiograph proves inconclusive, the lung function test is unremarkable and cough is the only presenting symptom, it will always be difficult to establish the diagnosis. The most common causes in the literature [5, 6, 8–10] are shown in table 4.
Learning points 1
Distinguish between acute (≤8 weeks duration) and chronic cough (>8 weeks); history and physical exam are usually adequate for diagnostic workup of patients with acute cough.
Chronic cough patients need immediate further diagnostic measures (usually chest radiography and spirometry ). If they prove inconclusive, check for upper airway disease, variant asthma or gastro-oesophageal reflux disease.
What are common causes of acute cough?
Enhancing the cough nerve sensitivity following diseases can cause acute cough.
Acute, self limiting viral infections of the upper and lower airways
Common cold is the most common cause of cough and usually subsides spontaneously, in otherwise healthy persons, after 2–3 weeks .
Upper airways allergic disease
Hay fever, and intermittent or persistent allergic rhinitis, often in combination with sinusitis, conjunctivitis, pharyngitis and laryngitis, can also trigger acute cough. Itchy eyes and throat are usually characteristic of these diseases .
Intermittent asthma, either allergic or due to infection, can cause acute cough.
Aspiration of a foreign body, most commonly in 1–3-yr-old children, as well as in elderly, fragile patients, triggers acute cough with expectoration of the foreign body, or permanent bronchial obstruction with consecutive chronic cough.
Acute inhalative intoxication
Workplace accidents, fires, and solvent or glue sniffing can lead to a toxic lung oedema, acute interstitial pneumonia and bronchiolitis with re-emergence of cough, often after a discomfort- and cough-free interval of 6–48 h. Immediate high-dose inhaled corticosteroid treatment should be initiated (≤100 puffs in 24 h).
Post-infectious cough persists >3 weeks after an acute, often viral airway infection and resolves after <8 weeks. Epithelial damage after B. pertussis or M. pneumoniae infection, or a transient increase in bronchial hyperresponsiveness (BHR), later subsiding spontaneously, are responsible for post-infectious cough. In the latter case, a short course of asthma treatment (inhaled corticosteroids or β2-adrenergics ) is effective.
Persistent BHR with consequent chronic cough without airflow obstruction is described as cough-variant asthma (see later). In this case, long-term inhaled corticosteroid treatment is required.
Pneumonia should be considered as a cause of acute cough.
Pleurisy is also a possible cause of acute cough.
50% of patients with acute pulmonary embolism present with a cough .
All forms of pneumothorax can be accompanied by a dry cough.
Acute heart failure with pulmonary congestion
Acute left heart failure (including lung oedema) can trigger both cough and bronchial obstruction [14, 15]. Bradycardia associated with acute emerging atrioventricular (AV) block II–III can greatly reduce stroke volume, eliciting pulmonary congestion and cough. Atrial premature beats can elicit acute cough .
Learning points 2
In otherwise healthy persons, common cold-related cough is self-limiting, usually lasting 2–3 weeks. OTC medication is appropriate. Antibiotics do not influence the natural history of this disease.
In the case of acute inhalative intoxication, immediately prescribe high-dose inhaled corticosteroid (≤100 puffs in 24 h).
Breathlessness, palpitation and acute cough are indicative of left heart failure, AV block and/or pulmonary embolism.
What are the causes of chronic cough?
Enhancing cough nerve sensitivity following diseases can cause chronic cough.
The World Health Organization defines chronic (nonobstructive) bronchitis as the presence of cough and phlegm on most days over a period of ≥3 months during two consecutive years without other causes. Many patients suffering from chronic cough meet these criteria. From a therapeutic point of view, this diagnosis is only useful if an obvious cause (i.e. smoking or work-related exposures) can be identified, cessation is possible and other causes of chronic cough have thoroughly been excluded. Though long-standing smokers frequently suffer from, but rarely complain of, cough and phlegm, chronic bronchitis is seldom a reason to attend a cough clinic.
Chronic obstructive pulmonary disease
By definition, patients with the chronic obstructive bronchitis phenotype of chronic obstructive pulmonary disease (COPD) are suffering from cough and phlegm. Chronic cough is a common symptom of COPD. In contrast to simple chronic bronchitis (see earlier), not fully reversible airflow limitation, as measured by spirometry, is the condition sine qua non for the diagnosis of COPD.
Asthma and other eosinophilic respiratory disease
Asthma is often responsible for chronic cough . Dry cough can elicit or worsen an asthma attack.
Cough-variant asthma is characterised by dry cough and BHR. Wheezing, dyspnoea and bronchial obstruction are absent. Chronic cough with proven BHR can only be confirmed as variant asthma if asthma treatment (inhaled corticosteroids or β2-adrenergics) eliminates the cough [10, 18–24].
Cough is the most common presenting symptom of lung tumours . If a patient presenting with chronic cough (i.e. lasting >8 weeks) is not taking an angiotensin-converting enzyme (ACE) inhibitor, chest radiography should be performed immediately. Furthermore, in order to exclude a lung tumour, each patient with unexplained chronic cough should have a bronchoscopy at the end of the diagnostic algorithm (fig. 2).
Cough due to upper airway disease
Chronic rhinitis and sinusitis (rhino-sinusitis)  are often associated with posterior (i.e. postnasal drip) and/or anterior mucopurulent drainage, nasal stuffiness, facial pain, pressure, and/or fullness and decreased sense of smell. Chronic rhino-sinusitis can occur with or without nasal polyposis and allergic fungal rhinosinusitis, respectively.
Vocal cord dysfunction
Recurrent voluntary inspiratory (sometimes also expiratory) adduction of the vocal cords can elicit throat clearing, dry cough, wheezing and dyspnoea. Vocal cord dysfunction can mimic asthma and often affects younger women .
Gastro-oesophageal reflux disease
Cough is triggered either by reflex, through reflux to the pharynx and larynx (laryngo-pharyngeal reflux), or microaspirations  of aerosolised gastric juice. Cough due to reflux can occur with or without heartburn  and does not necessarily coincide with reflux oesophagitis (nonerosive reflux disease). Thus, the gold standard of the reflux diagnosis is a triple-sensor, 24-h pH probe and impedance pH probe. The latter allows diagnosis of both acid and weakly acid reflux. Thus, aerosolised gastric juice, which probably plays a crucial role, can be measured directly by pharyngeal pH probe. Since pH probes are of limited sensitivity, frequently not available and poorly tolerated, high dose (2×40 mg) proton pump inhibitor treatment over the course of ≤3 months can be carried out as an alternative, thereby confirming or excluding the diagnosis of reflux cough  (fig. 2). However, unequivocal evidence for the efficacy of acid suppression from randomised controlled trials is lacking. In certain cases, surgical treatment (fundoplicatio) can be performed [34, 35], but no evidence-based selection criteria for surgery are yet available.
Approximately 10% of women and 5% of men cough while taking ACE inhibitor medication , which increases cough reflex sensitivity. The therapeutic (antihypertensive, cardiac or nephroprotective) effects of an ACE inhibitor treatment can be replaced by angiotensin II receptor antagonists, which do not cause cough more frequently than placebo. For other drugs inducing cough, information is available at www.pneumotox.com
In adults, pertussis is a rare cause of chronic cough, but has been described even without a preceding phase of acute infection. In particular, patients with recent contact with persons suffering from acute whooping cough should be checked for antibodies. However, interpretation of the results is difficult. After the acute exudative phase of infection (taking ≤10 days) a direct culture of Bordetella is no longer possible and antibiotics will have no effect on cough or on the natural history of the infection.
Chronic cough is a typical symptom of tuberculosis, and was one of its key diagnostic criteria in the pre-radiography era.
Chronic cough due to heart disease
Aside from chronic left heart failure (cough generally occurs upon physical exertion or prone position), AV block II–III, endocarditis  and cardiac arrhythmia [38, 39] can cause pulmonary congestion-related chronic cough. Moreover, chronic cough as a side-effect of cardiac drugs, including ACE inhibitors, β-blockers (in patients with BHR) and amiodarone (eliciting alveolitis) has to be considered.
Diffuse parenchymatous lung disease and systemic diseases with diffuse lung involvement
In addition to dyspnoea, dry cough is the most common symptom of diffuse parenchymatous lung disease (DPLD). Some forms of diffuse parenchymatous lung disease cause cough at such an early stage that the DPLD can be missed by conventional chest radiography. Thus, an apparently “normal” chest radiograph and spirometry do not rule out early lung disease with cough. A high-resolution computed tomography (HRCT) scan can establish the diagnosis.
Most systemic autoimmune disease can develop lung involvement and cause cough (e.g. Sjögren’s syndrome, systemic lupus erythematosus, rheumatoid arthritis, systemic sclerosis and vasculitides). Chronic cough in DPLD, however, can also be due to the treatment or infection.⇓]
Cough due to inhalative events
Chronic cough is caused if an aspirated foreign body becomes trapped in the bronchial system (usually in children 1–3 yrs of age), or due to chronic recurrent aspiration of food (liquids) resulting from dysphagia in underlying neurological conditions (e.g. bulbar paralysis, Parkinson disease  and myasthenia gravis). Other causes include tracheal−oesophageal fistula, malformations, neck dissection (head and neck cancer) and regurgitation in heavy gastro-oesophageal reflux disease
Reactive airways dysfunction syndrome (RADS) occurs following short-term, intense inhalation of vapours, smoke or gases  (usually due to accidents in the workplace) and often develops into difficult asthma.
Non-cystic fibrosis bronchiectasis and tracheobronchomalacia
Bronchiectasis can be missed on chest radiography. The gold standard for diagnosis is HRCT. Bronchiectasis usually causes productive cough with voluminous secretion, and often haemoptysis.
Cystic fibrosis is an autosomal recessive inherited disease. Abortive forms can manifest in adulthood for the first time through cough, bronchial infections and bronchiectasis.
Isolated orphan airways disease
Isolated orphan airways disease usually emerges in patients >40 yrs of age. It can lead to expiratory bronchial collapse and irreversible central obstruction of the airways. Coughing is frequently the main symptom (table 6).
Sleep apnoea patients often complain of chronic cough .
Psychogenic (habit) cough
By definition, the sensitivity of the cough reflex is not increased in patients with psychogenic cough, but this is difficult to measure reliably. There is always a risk of misdiagnosis of multicausal or idiopathic cough as psychogenic cough.
Chronic idiopathic cough
Despite extensive diagnostic procedures, underlying causes (e.g. reflux, asthma, etc.) of the increased sensitivity of the cough reflex (as measured by standardised capsaicin or citric acid testing) cannot be determined in ≤18% of patients with chronic persistent cough (females:males = 2:1) .
Learning points 3
Cough-variant asthma with BHR or eosinophilic bronchitis are responsive to inhaled corticosteroids.
If gastro-oesophageal reflux-related cough is suspected, evaluate double-dose proton pump inhibitor treatment for a period of 2–3 months to alleviate cough.
If a patient with chronic cough is using an ACE inhibitor, stop or replace their treatment for 3 weeks before starting further diagnostic workup (an exception from the rule), even if other causes for the cough are suspected.
Diagnosis of cough
Applying diagnostic algorithms frequently allows for a provisional diagnosis, which must be confirmed by successful treatment. Failure can, therefore, require continued investigation based on the algorithm. Multicausal cough requiring combination treatment also has to be considered. Figures 1 and 2 show algorithms for the diagnosis of acute and chronic cough, respectively.
Symptomatic treatment of cough
Causal treatment should always be sought. However, if this approach is impossible (e.g. acute viral respiratory infection) or would only prove effective in a delayed manner (e.g. tuberculosis), symptomatic treatment can be considered instead of, or complementary to, causal treatment of cough. Symptomatic treatment targets one or several of the five parts of the cough reflex arc. Effects can be protussive (increasing cough and expectoration) or antitussive.
Physiotherapy of cough
Despite being clinical routine in both hospital and outpatient care , as well as in rehabilitation, evidence for the efficacy of physiotherapy for cough is lacking. Physiotherapy aims to: 1) increase expectoration using effective coughing techniques for patients with productive but ineffective cough; 2) suppress voluntarily nonproductive cough; and 3) instruct patient in the use of physiotherapeutic equipment improving expectoration, such as Acapella® (DHD Healthcare, Wampsville, NY, USA), Flutter® (Desitin/Scandipharm VarioRaw SA, Birmingham, AL, USA) and RC Cornet® (BoniCur, Eastcote, UK).
Expectorants reduce irritation of the cough receptors by accumulated mucus through “coughing up”, and represent the most common medication used for respiratory diseases in Germany (e.g. ambroxol and N-acetylcysteine). Because of the lack of appropriate methods, effectiveness is difficult to assess. Regarding relative effectiveness of different expectorants, conflicting or inconsistent evidence exists throughout the published literature . Symptomatic use of expectorants is recommended to ease cough in cases with production of viscous secretions (COPD and bronchiectasis). Many patients also report positive subjective effectiveness using self-medication for acute bronchitis.
In cystic fibrosis bronchiectasis, inhaled dornase-α eases cough
Inhalative anticholinergics (i.e. ipratropium and tiotropium) are thought to reduce mucus production; however, their antitussive effect is not consistent .
Theophylline and β2-adrenergics do increase mucociliary clearance, but are not effective relieving cough.
By “coating” cough receptors in the throat, demulcents are thought to have an antitussive effect. Cough syrups, lozenges and drops, and honey share sugar as a common ingredient. Effectiveness, if any, is limited in time to the contact of the sugar with the receptor, which is usually 20–30 min.
Systemic α-adrenergics for nasal decongestion are popular in the US, but are virtually not in use in Germany and other parts of Europe. Fixed combinations with older, anticholinergic and central effective antihistamines (i.e. clorpheniramine or dexbrompheniramine) are not readily available in Europe. Moreover, evidence for their efficacy from randomised controlled trials is lacking.
Antibiotics are only effective against cough caused by a bacterial infection, characterised by purulent phlegm (i.e. suppurative bronchitis, bronchiectasis, exacerbation of COPD, purulent rhinitis and sinusitis). Antibiotics are not indicated in acute bronchitis.
Inhalative and nasal corticosteroids, and oral leukotriene antagonists alleviate cough in asthma, eosinophilic bronchitis, post-infectious cough due to BHR, and rhinitis.
Local anaesthetics disable electrophysiological activity in the receptors and afferent nerves (e.g. during bronchoscopy). They are increasingly used off-label for idiopathic cough and in palliative medicine .
Drugs affecting central mechanism for cough (antitussives) include systemically applied morphine or codeine, as well as natural and synthetic derivatives (i.e. dextromethorphan, dihydrocodeine, noscapine and pentoxyverin). Some nonaddictive herbal remedies (thyme, ribwort and sundew) claim central antitussive properties, though this is not proven by clinical studies. Opiates are recommended for effective symptomatic treatment of dry debilitating cough . They have limited efficacy in the treatment of cough resulting from common cold .
Learning points 4
First, consider causal treatment of the patient with cough if possible. If not, prescribe symptomatic treatment or the onset of the effect is delayed.
Consider physiotherapy and/or physiotherapeutic equipment for troublesome productive and nonproductive cough with or without bronchiectasis, even if there is no or only low-grade evidence for efficacy.
If there is a positive patient-reported effect, consider expectorants in bronchiectasis and COPD patients.
Chronic idiopathic cough and cough in palliative medicine: consider off-label treatments (morphine or local anesthetics, such as lidocaine or low-dose amitryptiline).
- ©ERS 2010