Published: March 2014
Cough is the most common single reason for primary care physician visits and is a common source of referrals to pulmonologists.1 In otherwise healthy persons, self-limited cough can occur as part of local irritation due to a viral rhinitis or respiratory infection. Chronic or persistent cough, defined as lasting more than 8 weeks, is usually not caused by a life-threatening disorder. However, the frequency of this complaint, its effect on quality of life, and concern about a serious underlying cause make chronic cough an important problem.
Because cough is a symptom that is a common pathway for a variety of diverse conditions and has no reliable objective monitoring tools, physicians' management of cough is actually quite variable. Clearly, a myriad of pulmonary parenchymal and airway diseases (e.g., interstitial lung disease and bronchiectasis) which present with abnormal chest imaging also have cough as one of the presenting symptoms. A common situation that clinicians face is the presence of a cough of unclear cause in the setting of a normal chest x-ray and normal spirometry, which is the focus of this section. A number of studies suggest that in addition to a history and physical, a systematic diagnostic approach including a chest x-ray, spirometry, bronchoprovocation study, sinus imaging, and esophageal pH monitoring yield a specific diagnosis in the vast majority of patients (>95%) with chronic cough.2 However, routinely obtaining all these tests is cumbersome and expensive in clinical practice, and some might not be readily available. Thus, it remains unclear whether most patients should undergo a trial of empirical therapy (either sequential or concurrent) or an aggressive and targeted diagnostic evaluation. In clinical practice, it is likely that these approaches are used in conjunction.
For a symptom such as cough, simply excluding certain serious causes can go a long way to reassuring the patient as well as the clinician and allowing a period of observation. An optimal, cost-effective approach to the management of chronic cough remains controversial. Because chronic cough is usually due to a benign cause, we recommend a stepwise approach employing empirical therapy targeted at the most common diagnoses, without extensive initial diagnostic testing (Figure 1). In some cases, a more aggressive approach may be necessary. For instance, in patients with a normal chest x-ray who have persistent cough and hemoptysis after antibiotic treatment, a fiberoptic bronchoscopic examination may be indicated to exclude endobronchial malignancy. Similarly, new cough or a persistent change in the character of cough in a patient with heavy smoking history may necessitate airway examination.
Cough is a protective physiologic reflex that augments mucociliary clearance of airway secretions. The cough reflex is characterized by the generation of high intrathoracic pressures against a closed glottis, followed by forceful expulsion of air and secretions on glottic opening. Intrathoracic pressures of up to 300 mm Hg and expiratory velocities approaching 500 miles per hour may be achieved.2 These high velocities serve to tear off mucus from the airway walls so the cough can effectively expel droplets into the air.3 Although it often serves a protective role, cough can become harmful when it is excessive and nonproductive. The high intrathoracic pressures and velocities generated during vigorous coughing produce hemodynamic changes analogous to chest compressions.4
Although physiologic cough has a characteristic acoustic profile and is universally recognizable, there is no clinical test that can document and confirm the presence of cough. This has hampered progress in the study of cough, and most of our understanding of the neuronal regulation of cough is derived from animal studies. The role of vagal afferent nerves in initiating the cough reflex is undisputed; however, the function of the various afferent neuronal subtypes in the regulation of cough is not well understood. The symptom of cough involves a reflex arc originating in peripheral cough receptors. Afferent nerves are most concentrated in the epithelia of the upper and lower respiratory tracts, but they are also located in the external auditory meatus, tympanic membrane, esophagus, stomach, pericardium, and diaphragm. For example, in a small percentage of patients, mechanical stimulation of the ear can evoke cough by stimulating afferent nerves carried by the auricular branch of the vagus nerve.5
Rapidly adapting receptors and C fibers are two afferent nerve subtypes that have been extensively studied and are theorized to play important roles in the regulation of cough. Rapidly adapting receptors are activated by mechanical forces such as lung inflation and deflation, whereas C fibers are much more sensitive to chemical stimuli, primarily bradykinin. Signals from the receptors are carried by vagal afferents to a medullary cough center, which then trigger cough activation via efferents mediated by the vagal, phrenic, and spinal motor nerves. Irritation anywhere along the reflex arc by a disease process can cause cough. The exact role of each of the various subtypes in the mediation of cough remains to be fully elucidated.5
Cough is the single most common symptom prompting outpatient medical visits in the United States.6 Persistent cough is a common reason for referral to a pulmonologist or allergist. The population point-prevalence of chronic cough depends on smoking status, ranging from 5% to 40%. Although cough can cause a variety of anatomic and physiologic complications, 98% of patients in one series listed the suspicion of underlying disease to be the major factor prompting them to seek medical attention. The aggregate cost of treatment for cough exceeds several billion dollars in the United States alone. This estimate does not include resources expended for repeated diagnostic studies.
Most patients seek medical attention because of complications of cough, either psychological or physical. The most common complications include feeling that something is wrong (98%), exhaustion (57%), feeling self-conscious (55%), insomnia (45%), lifestyle change (45%), musculoskeletal pain (45%), hoarseness (43%), excessive perspiration (42%), and urinary incontinence (39%).2 A host of other physiologic symptoms occur occasionally due to the high intrathoracic and intra-abdominal pressures achieved. The most prevalent symptoms include cough syncope, cardiac dysrhythmias, headache, subconjuctival hemorrhage, inguinal herniation, and gastroesophageal reflux. It is important to elicit which specific cough-related symptoms are bothersome for the patient as a guide to the pace and scope of diagnostic testing.
Although somewhat arbitrary, it is useful to classify cough by duration as a means of narrowing the differential diagnosis. Acute cough has been defined as one with a duration of less than 3 weeks at presentation.1 Some clinicians have proposed a category of subacute cough, with a duration of 3 to 8 weeks. Although this classification system is widely used and accepted by most experts, there are no case series assessing the relative frequency of causes in either the acute or subacute categories. Chronic cough is defined as lasting longer than 8 weeks. Most of the attention by clinicians is devoted to chronic or persistent cough because this is the variety that usually prompts patients to seek medical care.
The most common cause of acute cough, as defined by a time course of less than 3 weeks, is considered to be the common cold. Although there are no case series addressing the frequency of the common cold as a cause of acute cough, epidemiologic data support the relation. Acute and subacute cough are, by definition, self-limited. Cough that lasts longer than 3 weeks but does not become chronic falls into the subacute category. Postinfectious cough due to irritation of cough receptors accounts for most of these cases. Most commonly, persistent bronchial or sinonasal inflammation caused by a preceding viral upper respiratory infection is the culprit.
Acute bronchitis, as the term implies, is a lower respiratory tract syndrome and another common source of acute cough. It manifests as a persistent cough, with or without sputum production, in patients with a normal chest radiograph. Although it is much less prevalent than the common cold, acute bronchitis is the most common diagnosis given to patients presenting to a physician with acute cough. It is caused by a respiratory virus more than 90% of the time. Viral cultures and serologic assays are not routinely ordered; hence, the organism responsible is rarely identified.
Distinction between the common cold and acute bronchitis is often impossible to make and usually clinically irrelevant in otherwise healthy persons. Bacterial infections are implicated in less than 10% of cases of acute bronchitis. Despite this, reports have indicated that more than two thirds of patients receive antibiotic therapy for acute bronchitis. Elderly patients are more likely to receive broad-spectrum antibiotics, and nine out of ten smokers receive antibiotics despite the lack of evidence that smoking itself is a risk factor for bacterial infection. Treatment of self-limited bronchitis with antibiotics is often fueled by public expectation. In this circumstance, the descriptors a physician uses can make a difference. Patients are less likely to feel that antibiotics are necessary when the term "chest cold" is used as opposed to "bronchitis."7
A common diagnostic challenge encountered in the outpatient setting is determining the need for a chest radiograph. Pneumonia can also manifest with acute cough, and its manifestation can be more subtle in the elderly. Prospective studies have shown that the history and physical examination alone can predict the likelihood of pneumonia and therefore the need for a chest radiograph. A diagnosis of pneumonia is unlikely in the absence of tachycardia, tachypnea, fever, and evidence of focal consolidation on chest examination. The presence of purulent sputum does not distinguish between acute bronchitis and pneumonia.7
The most common causes of acute cough are listed in Table 1. In the presence of a compatible history and examination, further diagnostic testing is usually not necessary. Underrecognized causes of acute or subacute cough include pertussis and mycoplasma infection. Rarely, life-threatening illnesses manifest primarily with acute cough. Examples include pulmonary embolus, cardiogenic pulmonary edema, and pneumonia. The mainstay of treatment includes nonspecific antitussive therapy. Newer generation nonsedating antihistamines are not as effective in reducing cough compared with the older, first-generation antihistamine-plus-decongestant preparations. Bronchodilators are not recommended as adjunctive treatment unless baseline airflow obstruction is present. Expectorants and mucolytics, although commonly used, have not been proved beneficial.8
|Difficult to distinguish from acute sinusitis|
|Acute bronchitis||Antitussives||Antibiotics are not routinely indicated|
|Treatment duration is 2-3 wk|
|Allergic or irritant rhinitis||Irritant avoidance
|Nasal corticosteroids are useful for maintenance therapy|
|Acute exacerbation of COPD||β Agonist or anticholinergic MDI with spacer
|Pertussis||Macrolides (if started within 8 days of symptom onset) TMP-SMX (macrolide allergy)||Diagnosis is difficult (serologic or nasopharyngeal culture)|
COPD, chronic obstructive pulmonary disease; MDI, metered-dose inhaler; TMP-SMX, trimethoprim-sulfamethoxazole.
Determining the cause of chronic cough is often very challenging. As a general principle, the positive and negative predictive values of tests for the cause of chronic cough are poor, implying that attribution of cause depends on a response to specific therapy. Additionally, in 18% to 62% of patients, there are two significant causes, and in up to 42% there are three.9–11 In studies from cough clinics, the data suggest that there is little or no diagnostic value in descriptive features such as cough frequency, character, or sputum production. In addition, the most common causes of chronic cough can manifest with cough as the sole presenting symptom.
Several prospective studies have attempted to delineate the causes of chronic persistent cough and have reached the same conclusion. In the absence of smoking and angiotensin-converting enzyme (ACE) inhibitors, upper airway cough syndrome (UACS), asthma, and gastroesophageal reflux disease (GERD) are the most common causes of chronic cough in patients with a normal chest radiograph.12 The same triad applies in children and the elderly.2,13 Outside the United States, nonasthmatic eosinophilic bronchitis (NAEB) is a recognized cause of chronic cough, accounting for up to one third of cases in some series. Even in immunocompromised hosts and areas where tuberculosis is endemic, UACS, asthma, GERD, and NAEB are the most common causes reported. Chronic bronchitis, usually resulting from cigarette smoking, is believed to be the most common overall cause of chronic cough, but most smokers with cough typically do not seek medical attention. All of these data are derived from studies of patients referred to pulmonologists or cough clinics; however, it is likely that a similar spectrum of causes accounts for most patients seen by primary care providers.
In January of 2006, the American College of Chest Physicians published a guideline on the diagnosis and management of cough. The review on postnasal drip (PND) addressed the unclear relation of PND to cough. Given the lack of understanding about whether PND is the cause of cough or a consequence of airway inflammation that is producing cough, the term upper airway cough syndrome was adopted as a more accurate descriptor.
UACS is the single most common cause of chronic cough, accounting for 8% to 87% of cases, either exclusively or in combination with other factors.2,13 Categories of UACS-induced cough include allergic rhinitis (seasonal or perennial), perennial nonallergic rhinitis (vasomotor or nonallergic rhinitis with eosinophilia), postinfectious UACS, bacterial sinusitis, allergic fungal rhinitis, occupational rhinitis, rhinitis medicamentosa, and pregnancy-associated rhinitis. All patients should undergo evaluation for exposure to common allergens or chemical irritants. An episodic rhinitis might suggest that an occupational factor is involved.14
The pathogenesis of cough in UACS involves stimulation of afferent receptors in the upper airway, rather than run-off of secretions into the lower airway. Associated symptoms include rhinorrhea, nasal congestion, a sensation of drainage or tickle in the oropharynx, and throat clearing. Physical examination can reveal nasal congestion or discharge, nasal mucosal bogginess, mucous in the oropharynx, or a cobblestone appearance of the oropharyngeal mucosa. The examination findings are nonspecific, however, and may be present in any of the other major causes of chronic cough. Although most patients have at least one symptom or sign, UACS can manifest as cough alone up to 20% of the time.12 Because it is highly prevalent and may be otherwise clinically unapparent, it is reasonable to employ a brief trial of empirical treatment in the approach to diagnosis. Ultimately, UACS is a syndrome without a clear definition, and its role in chronic cough is best proven by a response to therapy. Because the recommended therapy (older generation antihistaminics) can suppress the cough reflex centrally and peripherally, response to therapy may not necessarily confirm that the diagnosis is accurate.
The presence of copious sputum is associated with an increased likelihood of chronic sinusitis, but neither the clinical examination nor historical features reliably differentiate it from other causes of UACS.11,13 Excessive sputum production, defined as greater than 30 mL per day, can be a manifestation of asthma, GERD, and bronchiectasis. Patients with chronic sinusitis can present with a nonproductive cough as their sole symptom.
Among patients with chronic cough, up to 38% have some radiologic sinus abnormality.15 Thus, a finding of sinus mucosal thickening on radiographs has only a 29% to 81% positive predictive value (PPV) for chronic sinusitis that is responsible for cough.2,15 Additionally, mucosal thickening is seen in bacterial and viral infections and cannot be used to distinguish between the two. The presence of air-fluid levels on four-view sinus x-rays is more specific for diagnosing sinusitis, and one report documented a 100% PPV.15 Due to the poor PPV, CT scan of the sinuses is not encouraged routinely for the diagnosis of UACS1.
In most series, asthma is the second most common cause of chronic cough in adult nonsmokers, present in 14% to 55% of cases.16,17 Historical features of wheezing, chest tightness, or exertional dyspnea in response to triggers such as strong odors or perfumes, cold air, or allergens should suggest the diagnosis. However, the PPV of a suggestive history is only 56%.18
Cough is a feature in nearly all cases of asthma. In a subset of asthmatic patients, cough is the primary or sole symptom, a condition termed cough-variant asthma (CVA). CVA accounts for 6.5% to 57% of all asthmatic patients.2 Many of these patients later develop classic asthma symptoms.18
The physical examination and spirometry can be entirely normal in patients with chronic cough due to asthma. The most useful test to support the diagnosis is a bronchoprovocation test with inhaled methacholine. The methacholine challenge test (MCT) is highly sensitive, and a negative test virtually excludes the diagnosis of asthma. A 20% decrement in forced expiratory volume in 1 second (FEV1) after methacholine inhalation, although indicating bronchial hyperresponsiveness, can have a PPV as low as 74% for diagnosing the cause of cough.19 More recently, measurement of exhaled nitric oxide has proved useful in the evaluation of chronic cough, presumably by identifying cases due to CVA. When 30 ppb was taken as the diagnostic threshold, exhaled nitric oxide testing had a sensitivity and specificity of 75% and 87%, respectively.20 Frequently, however, proof that asthma is the inciting factor in chronic cough requires demonstration of a response to directed therapy, i.e., inhaled corticosteroids. The utility of using short-acting bronchodilators as empirical therapy for CVA is unknown.
GERD, alone or in combination with other conditions, accounts for up to 40% of cases of chronic cough around the world.16 It has been recognized as a contributor to cough with increasing frequency in observational studies; in some series it has surpassed other etiologies as a cause for chronic cough.2,16 Recognition, diagnosis, and treatment of GERD are among the most challenging of all cough etiologies, with a number of unclarified issues. Patients with GERD are commonly asymptomatic, and it can be assumed that the condition is therefore underrecognized.
In some cases, reflux-mediated irritation of laryngeal receptors or episodic microaspiration underlie GERD-induced cough. This condition has been termed laryngopharyngeal reflux. Nevertheless, GERD-induced cough is frequently attributable to a reflex loop involving vagal afferents in the distal esophagus, and proximal reflux is unnecessary in the pathogenesis.21 Reflux into the distal esophagus alone can stimulate the esophageal-bronchial cough reflex. It is important to recognize that nonacid reflux can cause chronic cough, and these patients do not respond to acid-suppression therapy. Coughing itself can induce reflux by an unknown mechanism potentially creating a vicious self-perpetuating cycle.22
Symptomatic heartburn occurs in only a minority (25%-50%) of patients.2 Other historical features such as exacerbation at night, in the supine position, or after eating do not reliably differentiate GERD-induced cough from other causes.16 GERD also commonly accompanies other causes of cough; up to 80% of asthmatic patients have abnormal 24-hour pH probe findings.2 Recurrent elevations in abdominal pressure can contribute to this phenomenon.
Ambulatory 24-hour esophageal pH monitoring is the most reliable test for GERD. It is important to include a temporal symptom log when conducting pH monitoring so as to document the causality of reflux events vis-à-vis cough. Often, close examination reveals that cough preceded the reflux event. A temporal symptom log can also substantiate GERD as a cause of cough even when pH probe scores lie within the normal range. In untreated patients, the pH probe carries a 90% to 100% sensitivity, but the PPV may be as low as 35% when using therapeutic response as a gold standard.2,23
The esophageal pH monitoring study will be negative in patients with nonacid reflux, and if this condition is suspected, barium esophagography may be helpful. Esophageal intraluminal impedance plethysmography may also be able to identify patients with nonacid reflux. However, there are no published data to suggest that outcomes are improved with this device.
Most patients with GERD do not have endoscopic findings that suggest acid reflux, such as esophagitis or Barrett's epithelium, and a normal endoscopy does not rule out GERD as a source of cough. It is reasonable to empirically treat patients who fit the clinical profile before testing. However, recent data indicate that empirical treatment of patients with chronic cough who have rare or no heartburn may not improve cough related quality of life or symptoms.24 Similar to other causes of chronic cough, a diagnosis is suggested only when the cough resolves with treatment.
A subset of patients has been recognized with increased sputum eosinophils in the absence of demonstrable bronchial hyperresponsiveness. These patients are typically nonsmokers and have a chronic cough that responds to inhaled corticosteroids. The frequency of this syndrome has been reported to be as high as 10% to 30% in patients with chronic cough who are referred to a specialist.25 Diagnosis is usually made by demonstration of induced sputum eosinophilia (>3%). If induced sputum cannot be obtained or is nondiagnostic, bronchoscopy with bronchial wash can provide useful information. The pathogenesis of NAEB is not well understood, although, similar to asthma, an inhaled allergen or environmental agent is theorized to play a role. Persistent eosinophilic airway inflammation can lead to progressive airflow obstruction, and NAEB may be implicated as a causative factor in chronic obstructive pulmonary disease (COPD).26 The frequency and significance of NAEB remain incompletely understood.
Chronic bronchitis is characterized by a productive cough on most days for 3 months in 2 consecutive years. It may be caused by irritant-induced inflammation or by the need to mobilize excessive secretions. Although chronic bronchitis is a common cause of cough in the population, it is present in only 5% of those seeking medical attention for cough.9,10 Cigarette smoke is the most common irritant associated with chronic bronchitis. There is a direct relation between the incidence of chronic bronchitis and the number of cigarettes smoked. As the airflow obstruction becomes more severe, the incidence of cough increases. Clinicians should also inquire about passive smoke exposure because it is linked to chronic productive cough. Occupational exposures or inflammatory bowel disease can also trigger this syndrome.27
The cause of cough in chronic bronchitis is multifactorial, with bronchial secretions and airway inflammation playing a primary role. Patients with airflow obstruction cannot generate an effective cough to clear secretions. Retained secretions can worsen cough and incite a self-perpetuating cycle. Acute respiratory infections, whether viral or bacterial, are a common cause of exacerbations in chronic bronchitis. Patients typically present with increased cough, sputum production, and worsening dyspnea.27
The mainstay of treatment is avoiding the offending agent(s). Cough completely resolves or significantly decreases in 90% of patients after smoking cessation. In contrast to acute bronchitis, exacerbation of chronic bronchitis should be treated with antibiotics. Particular vigilance is warranted when there is a change in the character of the cough or sputum, because this may be the manifesting feature of a superimposed bronchogenic carcinoma.
Airway inflammation can follow any respiratory tract infection; up to one quarter of such infections may be complicated by persistent cough.2 Increased cough receptor sensitivity or temporary bronchial hyperresponsiveness, probably related to epithelial damage, likely account for this phenomenon. Postinfectious cough is a diagnosis of exclusion, and by definition it eventually resolves, but its duration may be prolonged. For this reason, 8 weeks is a more clinically useful working definition of chronic cough than 3 weeks. Postinfectious cough is usually caused by respiratory viruses, Mycoplasma spp, Chlamydia pneumoniae (strain TWAR), or Bordatella pertussis. Vigorous coughing can stimulate esophageal reflux, which has been implicated as a cause of postinfectious cough.82
Pertussis, or whooping cough, is a severe and debilitating cough illness that can last for weeks to months. A common misconception is that pertussis is a disease of infants and children. Active pertussis infection in adolescents and adults is an underrecognized cause of chronic cough. One investigation found a 21% incidence of pertussis in a group of patients with a cough duration of 2 weeks to 3 months.29 In recent years, the incidence of pertussis has increased, most likely due to a combination of waning immunity and increased detection of the illness. Patients experience periodic episodes of violent coughing, and post-tussive emesis is common. The inspiratory whooping sound typical in children is generally not heard in adults.28,30 Diagnosis of cough due to pertussis may be aided by demonstration of positive polymerase chain reaction in nasopharyngeal secretions in the acute setting and serum immunoglobulin G analysis later on in the course.31
ACE normally degrades proinflammatory mediators, such as bradykinins and substance P. Inhibition of this action lowers the threshold for cough sensitivity. Cough due to ACE inhibitors is a class effect and has been documented with all ACE inhibitors. Switching to another agent will not ameliorate the symptoms. Usually, cough begins within 1 week of starting the medication, but it can begin as late as 1 year later. Approximately 10% to 20% of patients develop cough, but in many it is not bothersome. Asthmatic patients are not at increased risk. Cough resolution may be delayed for up to 1 month after discontinuing the drug. Angiotensin II receptor blockers do not cause cough and therefore are useful therapeutic alternatives.
Chronic cough may be unexplained despite comprehensive workup and empirical treatment trials in a substantial number of patients. In referral centers, the prevalence has been reported to be as high as 42% of cases.32 A heightened sensitivity to cough challenge with capsaicin has been demonstrated in these patients. There is a frequent history of viral illness preceding the onset of chronic cough. The syndrome appears to have a predilection for females who have onset of cough around menopause. Interestingly, a high prevalence of vocal cord dysfunction has been reported in association with cough hypersensitivity syndrome.33 It has also been suggested that the otolaryngological disorders such as laryngeal sensory neuropathy overlap considerably with the cough hypersensitivity syndrome.34 Treatment is challenging but speech therapy may hold promise.35 Nebulized local anesthetics, gabapentin, and amitriptyline have also been used with variable results.
Psychosomatic cough is rare, especially in adults. It is a diagnosis of exclusion and should not be entertained until all other potential causes have been investigated, with therapeutic trials of adequate intensity and duration. There are no distinguishing historical features to reliably differentiate it from other etiologies.16 Habit cough is a syndrome of persistent, habitual throat clearing that might respond to biofeedback. In distinction, psychogenic cough usually implies an underlying psychiatric disorder. Most patients with psychogenic cough harbor an intractable fear of a serious underlying medical disease.
Disorders of any of the locations of cough receptors (external auditory canal, tracheobronchial tree, pleura, pericardium, diaphragm, esophagus, stomach) can cause chronic cough. Recently, obstructive sleep apnea and chronic tonsillar hypertrophy have been recognized as associations with chronic cough. In addition to ACE inhibitors, drugs such as sitagliptin, inhaled corticosteroids, topiramate, erythromycin, ribavirin and methotrexate have been associated with chronic cough. Beta blockers can aggravate CVA. Vitamin B12 deficiency has been associated with chronic cough via induction of cough reflex hypersensitivity.
The management of chronic cough typically involves some combination of simple screening studies (chest x-ray and spirometry), additional specific diagnostic studies (methacholine provocation, sinus imaging, or a pH probe), and empirical therapy for the three most common entities (rhinitis, asthma, GERD) (see Figure 1). Patients who lack specific clues by history and examination and who have a normal chest x-ray and spirogram represent the most common management dilemma for the clinician. A major decision involves the extent of specific diagnostic testing as opposed to trials of empirical therapy. Randomized clinical trials do not provide adequate guidance to help the clinician choose between these two strategies. The approach is usually negotiated with the patient, partly based on the level of subjective distress and on the level of exasperation by the patient and the clinician.
Pinpointing the cause of chronic cough is often difficult. Cough may be the sole manifestation of disorders such as asthma, GERD, or UACS, with a paucity of other historical features to suggest the correct diagnosis. Features of cough, such as timing, associated sputum production, and cough character (e.g., brassy), are not helpful in distinguishing causes.16 A high percentage of patients have two or more responsible causes. Finally, there are no diagnostic tests with a sufficiently high PPV to reliably implicate any particular cause of cough. Thus, diagnosis hinges on demonstration of a response to a specific therapy.
Evaluation and treatment using our algorithm (see Figure 1) assume that failure to remedy the cough using trials of empirical therapy will precipitate appropriate diagnostic testing. A significant fraction of nonresponders fail treatment due to inadequate intensity or duration of treatment. One study found that the diagnosis was correct in 14% of referred patients, but the treatment regimens were insufficient.10 In these cases, diagnostic testing facilitates appropriate narrowing and intensification of treatment. Because the overwhelming majority of patients have UACS, asthma, or GERD, it is crucial to assiduously investigate the roles of each before further investigations.
Common causes for diagnostic frustration include inadequate diagnosis and inadequate treatment. All tests used for evaluating chronic cough have a poor PPV. Over-reliance on historical features or cough characteristics can thwart accurate diagnosis because these are often misleading. The cause of cough may be otherwise clinically silent, and 18% to 62% of chronic coughs are due to two or more causes. Diagnostic testing that suggests an underlying cause does not ensure that cough is caused by that etiology. Inadequate treatment regimens are a common reason for failure to alleviate symptoms, and eradication of related symptoms (heartburn, nasal congestion) does not ensure that an underlying cause is sufficiently treated. Treatments themselves can frustrate the search. Central cough suppressants, such as dextromethorphan and codeine, should generally be avoided; they might serve as a temporizing measure but can distract from the search for a specific cause. Pressurized metered-dose inhalers (MDIs) can exacerbate CVA.
The American College of Chest Physicians and the British Thoracic Society advise obtaining a chest x-ray before starting empirical therapy.36 Although most investigators recommend chest x-ray early in the evaluation of chronic cough, only 4% to 11% of nonsmokers have culpable abnormalities.9,10 The yield of fiberoptic bronchoscopy is similarly low, and only 4% of patients with normal chest x-rays have endobronchial abnormalities. Even with a finding of endobronchial pathology, the PPV is only 50% to 89%.2 In assessing for the presence of uncommon causes of chronic cough, chest CT scanning has a relatively higher diagnostic yield (abnormal in up to 42% of patients with a normal chest radiograph) and should be performed before cardiac tests in the absence of cardiac symptoms.36
Some situations merit special consideration. Consider foreign body inhalation in patients who present with sudden-onset cough. Consider sinus imaging if a patient suspected of having rhinitis or UACS does not respond to empirical therapy; chronic sinusitis can be otherwise clinically silent.14 In patients who do not respond to empirical treatment for UACS and GERD and who have a normal chest x-ray, normal spirometry, and negative MCT, it is reasonable to empirically treat with a short course of inhaled corticosteroids. Resolution of cough would support the diagnosis of NAEB. Occupational and environmental exposures should be considered as exacerbating factors in all patients who present with cough. Indoor allergens, such as pets and dust mites, and secondhand exposure to cigarette smoke are common environmental causes that are easily modifiable.29
Initial therapy for most postnasal drainage syndromes should include an antihistamine-decongestant combination (Table 2). First- generation antihistamines, such as dexbrompheniramine maleate or azatadine maleate, have been demonstrated to be superior to second-generation (nonsedating) drugs, due to their additional anticholinergic activity.2 If drowsiness is problematic, therapy may be initiated with bedtime dosing before instituting twice-daily dosing. Second-generation antihistamines are useful primarily in allergic rhinitis syndromes. The role of bacteria in perpetuating chronic sinusitis is controversial, and treatment regimens are not well defined. Except for chronic sinusitis, in most patients symptoms respond within 1 week. Long-term use of topical decongestants should be discouraged to avoid rebound nasal congestion (rhinitis medicamentosa).
|Post-infectious||First-generation antihistamine plus decongestant combination||Dexbrompheniramine maleate 6 mg plus pseudoephedrine 120 mg bid||Nasal corticosteroids are useful for maintenance therapy
Ipratropium nasal MDI is also effective
|Allergic rhinitis||Second-generation antihistamine
|Loratidine 10 mg/day
Beclomethasone nasal 42 µg each nostril qid
|Cromolyn, azelastine, leukotrine inhibitors are also useful
Consider allergy testing
|Non-allergic rhinopathy (vasomotor rhinitis)||Anticholinergic nasal spray||Ipratropium 0.06% 2 puffs/nostril 4-6 times/day||Intranasal glucocorticoids and/or intranasal antihistamine if congestion, sneezing present|
|Chronic sinusitis||First-generation antihistamine plus decongestant (3 wk)
Nasal decongestant (5 days)
Appropriate antibiotics (≥3 wk)
Pseudoephedrine as above
Oxymetazoline 2 sprays bid
|Use antibiotic active against H. influenzae, S. pneumoniae, oral anaerobes
Nasal corticosteroids after cough resolves
MDI, metered-dose inhaler.
Initial empirical therapy with an antihistamine-decongestant has been shown to be effective in a large percentage of patients who present with chronic cough. This reflects the prevalence of UACS among patients with chronic cough.19
Inhaled corticosteroids are generally required to optimize therapy for most patients. Using a pressurized MDI can aggravate cough which may be alleviated by the addition of a spacer. Mast cell stabilizers, such as cromolyn sodium, are also effective. A few patients require oral corticosteroids for symptom control. Zafirlukast, a leukotriene inhibitor, has been shown to have an antitussive effect in patients with CVA.
The presence of bronchial hyperresponsiveness should be demonstrated by provocation testing or reliable history before initiating oral steroids. Caution is warranted when interpreting the results of empirical therapeutic success with asthma therapy; NAEB and postinfectious cough might respond similarly. For this reason, we recommend at least one attempt to taper therapy. Steroids should also be tapered in patients whose MCT was obtained in the setting of a postinfectious cough.
Treatment of GERD should begin with maximal therapy. Patients should avoid reflux-predisposing foods (fatty foods, chocolate, caffeine, alcohol), give up tobacco, elevate the head of the bed, and not eat within 2 or 3 hours before lying down. These are important lifestyle adaptations that can significantly diminish cough. Proton pump inhibitors (PPIs) should be used as the mainstay of medical therapy. In nonsmokers with a normal chest x-ray in whom asthma, UACS, and NAEB have been ruled out, empirical treatment has been shown to be cost effective.22 We recommend commencing treatment with twice-daily PPIs, dosed half to 1 hour before the morning and evening meals; later, therapy may be de-escalated if proven effective. Onset of response can take up to 3 months, with maximal effects of treatment delayed for up to 6 months.2 Response rates of 70% to 100% have been reported when the end point is improvement in cough.2 However, the true response rate is unknown, because therapeutic response is the only indication; there is no objective gold standard to prove GERD is the cause of cough. The role of nonacid reflux in cough pathogenesis is also currently unknown, but it probably accounts for some of the nonresponders.
Ambulatory pH monitoring should be considered in patients who fail maximal therapy. Unfortunately, this test is usually only available in larger academic centers. Treatment response in patients with GERD can be prolonged and even take up to 6 months. Intensive treatment for at least several months should be attempted before referral for further diagnostic testing. Documentation of persistent symptomatic acid reflux should prompt referral to a specialist. Surgical correction with esophageal fundoplication is controversial and can result in significant comorbidity.
Inhaled corticosteroids are usually effective, but a few patients require a short course of oral steroid therapy.25 A 2- to 3-week course of prednisone at 30 mg/day is sufficient; a small percentage of patients require maintenance therapy, generally with inhaled corticosteroids.
Smoking cessation markedly reduces symptoms in more than 50% of patients within a month. Cough frequency and sputum production are most effectively decreased by ipratropium MDI (2 puffs qid). Mucolytics are not helpful, and the symptomatic benefit of ICS has not been demonstrated.
Postinfectious cough usually responds to treatment analogous to asthma. Alternatively, ipratropium has proven efficacy. Ongoing infections should be treated with appropriate antibiotic therapy (see Table 1).
Nonspecific cough suppression becomes necessary when the symptom impacts quality of life significantly or when it is due to a disease process without effective treatment (e.g., idiopathic pulmonary fibrosis). Narcotics (morphine, codeine and dextromethorphan) have traditionally been used for cough suppression but have well known and prohibitive side effects with chronic use. Benzonatate is a valuable option for cough suppression; however, side-effect profile includes seizures and cardiac arrhythmias. Effective pharmacotherapy for cough suppression remains an unmet need. Recent introduction of central cough suppressants such as cloperastin37 and peripheral cough suppressants such as levodrpropizine38 and moguisteine39 have been welcome additions although not yet available in the United States. The discovery of upregulation of the transient receptor potential (TRP) nociceptors in patients with cough reflex hypersensitivity promises new peripherally active agents (TRP receptor antagonists) in the treatment of stubborn cough.
Evaluation of cough depends on a systematic review of potential causes. The balance between empiricism and testing is often dictated by the patient's clinical course and quality of life. In most cases, empirical therapy targeted at the most common diagnoses leads to significant improvement or resolution of cough, obviating the need for extensive diagnostic testing. A chest radiograph and spirometry are recommended before referring a patient to a specialist.
The cause of chronic cough may be otherwise clinically silent, and it is not uncommon for there to be more than one contributing factor. Ultimately, the underlying cause may be inferred by demonstrating a response to therapy. Complete resolution of chronic cough can take weeks to months, and both the physician and patient need to set realistic treatment goals. Inadequate length of treatment, insufficient medication intensity, and patient nonadherence are common reasons for treatment failure. For this reason, a combination of systematic empiricism and limited adjunctive diagnostic testing are appropriate for most patients.