Partners Asthma Center Grand Rounds

Barbara A. Cockrill, M.D.
Intractable Cough

Case Example

A 48 year-old office worker was referred for evaluation of a chronic cough. She had had a cough for many years, and in recent months it seemed to worsen. Her cough sounded barking-like and "juicy," but no phlegm was expectorated. She had some post-nasal drip that was being adequately treated; she denied any symptoms of esophageal reflux. She was a lifelong nonsmoker and did not take an angiotensin converting enxyme inhibitor.

Her cough had an enormous impact on her life. Due to embarrassment about her repetitive coughing, she stopped going to movies, eating out at restaurants, and attending parties.

Initial work-up of her cough included spirometry, chest radiograph, and chest CT scan, all normal. Methacholine challenge was likewise normal.

An esophageal pH probe study proved positive for esophageal reflux. However, intensive medical treatment for reflux disease (and for post-nasal drip) did not ameliorate her cough. She then underwent bronchoscopy, which was unremarkable (other than her troublesome cough during the procedure). She received empiric anti-asthmatic therapy, including a course of oral prednisone, without improvement.

The patient was referred for behavioral modification therapy with a psychiatrist, without benefit. A repeat pH probe study performed with the patient on anti-reflux therapy was again positive for esophageal reflux. Despite the absence of esophagitis on upper gastrointestinal endoscopy, she proceeded to surgery for esophageal reflux with a Nissen fundoplication. Despite initial improvement, her cough recurred unabated after about two weeks. (Case to be continued...)

This Grand Rounds presentation will cover the normal physiology of cough, some of the common syndromes causing cough with a normal chest X-ray, and non-specific cough suppressant therapies. The important "take-home" message is that even in a referral practice, common causes of cough are most common. Using a systematic approach one can almost always identify the cause or causes of cough, and most often (more than 90% of the time) one discovers one (or more) of a few common etiologies. Treatment failures prior to referral to a specialist usually reflect failure to recognize one of these common causes of cough or failure to treat it with sufficient intensity.

Cough accounts for more than 30 million visits per year to primary care physicians in the United States. It can cause sleep disturbance, incontinence, headache, musculoskeletal pain, vomiting, rib fracture, and syncope. Cough exists as a vital protective mechanism for the respiratory system; its absence causes even greater problems than its pathologic presence.

Normal Physiology

Four discrete steps are involved in a normal cough. The first is an inspiratory gasp. The deep inspiration stimulates pulmonary J receptors that augment the cough stimulus via the Herring Breuer reflex. This step is impaired — and cough is weak — in patients with inspiratory muscle weakness. Step two is the Valsalva maneuver: forced exhalation against apposed vocal cords. Patients with an open tracheostomy or paralyzed vocal cords have impaired cough because of their inability to raise intrathoracic pressure against a closed glottis.

The third step is the expiratory blast, as the vocal cords suddenly abduct. Expiratory muscle function is crucial for this step; expiratory muscle weakness (e.g., patients with spinal cord injuries above T6) causes ineffective cough. Finally, in step four there is a post-tussive prolonged inspiration.

Figure 1 shows diagrammatically some of the neural pathways involved in the normal cough mechanism. The cough center is located in the medulla and pons. Cough receptors in the nose are thought to be innervated by the trigeminal nerve. In the vocal cords they receive innervation from the superior laryngeal nerve. Below the vocal cords, other branches of the vagus nerve carry signals from cough receptors.

In general, the higher you are in the respiratory tract, the more the cough receptors respond to mechanical stimulation; the lower you are in the respiratory tract, the more the cough receptors respond to chemical stimulation.

The Common Causes of Chronic Cough

Dr. Richard Irwin and colleagues have reported their experience evaluating patients referred for chronic cough of at least three weeks' duration. Work-up included history, physical examination, and chest X-ray in all patients, then further testing as indicated. They found that in nearly three fourths of their patients, asthma and/or post-nasal drip was the cause of persistent cough. Other common causes included chronic bronchitis, gastroesophageal reflux disease (GERD), and post-infectious cough. Rarer causes were occasionally discovered, including bronchogenic cancer, sarcoidosis, congestive heart failure, and esophageal diverticulum.

A systematic approach to chronic cough, derived from the work of Dr. Irwin and his colleagues, is displayed in Figure 2 . History and physical examination may indicate a diagnosis, such as the onset of cough following a respiratory tract infection (post-infectious cough). Nocturnal cough suggests possible asthma, esophageal reflux, or congestive heart failure. Cough related to meals may indicate aspiration or regurgitation. An early morning productive cough is typical of chronic bronchitis and bronchiectasis. Seasonal cough or exercise-induced cough points to a diagnosis of asthma. Cough with hoarseness raises the possibility of laryngeal disease or esophageal reflux. Among medications causing cough, angiotensin converting enzyme inhibitors are the most common.

When no diagnosis is evident from patient interview and exam, a chest X-ray and spirometry should be obtained. A radiographic or spirometric abnormality will direct further work-up. If the chest X-ray and spirometry are normal, one needs to start thinking a little more creatively. Consider a full set of pulmonary function tests with lung volumes (to detect unsuspected restrictive pulmonary disease), a methacholine challenge (to evaluate for bronchial hyperresponsiveness), and esophageal pH probe (for GERD). An alternative approach utilizes a therapeutic trial of treatment for asthma or GERD. The relative risks and benefits of these two approaches have not been fully evaluated.

Bronchoscopy has a relatively low yield among patients with chronic cough and a normal chest X-ray. In a retrospective review by Dr. Robert Poe and colleagues, only one bronchoscopy proved diagnostic out of 51 procedures performed among 109 patients referred for evaluation of chronic cough. Nonetheless, bronchoscopy may occasionally be indicated. For example, in one series 17 of 46 patients with carcinoid tumors presented with chronic cough and clear chest radiographs. In another case series six patients with postoperative retained suture material had cough and unremarkable chest films. We tend to perform bronchoscopy after diagnostic testing or treatment trials have excluded all of the more common causes of cough.

Specific Cough Syndromes

Post-nasal drip is probably the most common cause of chronic cough. Patients generally experience a sensation of mucus dripping or a tickle in the back of their throats. On physical exam one finds "cobblestoning" of the mucosa of the posterior pharynx, representing mucous gland hyperplasia. Treatment depends on the cause of the rhinorrhea. Options include antihistamines (including the older antihistamines with greater anticholinergic effects), decongestants, nasal steroids, nasal ipratropium, and, when bacterial sinusitis is present, antibiotics.

It is said that as many as 25% of persons with asthma may present with cough as their sole symptom ("cough-variant asthma"). Spirometry is typically normal. The diagnosis can be made by identifying bronchial hyperresponsiveness on bronchoprovocation challenge or by resolution of the cough with treatment for asthma. It has been hypothesized — but not confirmed with endobronchial biopsies — that cough-variant asthma is a manifestation of more proximal airway inflammation, whereas more typical asthma with wheezing involves peripheral airways.

It is easy to understand why persons who have gastroesophageal reflux to the level of the pharynx with subsequent aspiration of gastric contents would manifest with cough. What about patients who have GERD but do not reflux to the level that might lead to aspiration or irritation of the larynx? An interesting study compared persons with GERD to normals in their cough-sensitivity to capsaicin (the chemical that makes hot peppers "hot"). Patients with GERD required a much lower dose of capsaicin to induce coughing than normals. Airway responsiveness to methacholine was normal in both groups.

In a study further investigating the mechanism by which GERD might cause cough, hydrochloric acid was instilled onto the distal esophagus in persons with GERD and cough and in normal controls. Acid instillation caused more cough among persons with GERD. This coughing could be attenuated by pretreatment with lidocaine instilled onto the esophagus and by ipratropium inhaled onto the airways, but not by ipratropium installation onto the esophagus. This study suggests a reflex mechanism involving afferent neural pathways in the esophagus and efferent pathways in the trachea or more distal airways.

It has been reported that as many as 75% of patients with cough due to GERD will not have symptoms of acid reflux or "heartburn." In an analysis of various testing procedures to establish a diagnosis of GERD, esophageal pH monitoring was found to be the most sensitive procedure compared to barium swallow and endoscopy. Treatment of GERD should begin with conservative measures such as elevating the head of the bed and dietary adjustments. Proton pump inhibitors such as omeprazole or lansoprazole are the most effective medical therapies. Relatively high doses of the proton pump inhibitors (e.g., omeprazole 20 mg twice daily) are needed for treatment of GERD-induced cough. Motility agents such as metoclopramide or cisapride are also effective and may provide additive benefit to the proton pump inhibitors. Surgery represents a final option. Surgical approaches include fundoplication and, among patients with delayed gastric emptying, pyloroplasty.

Post-infectious cough typically follows certain viral respiratory tract infections. Patients may have a component of bronchoconstriction and will show evidence for airway hyperresponsiveness on methacholine challenge testing. Treatment with bronchodilators and inhaled corticosteroids is often effective in relieving the cough; so, too, is the tincture of time. As Voltaire said, "The art of medicine is amusing the patient while nature cures the disease."

A special instance of post-infectious cough is pertussis or "whooping cough," caused by the bacteria Bordetella pertussis or B. parapertussis. Most infants in the United States are vaccinated against pertussis, but immunity wanes after about 12 years. Safe, acellular vaccines are now available. Although reasonable, especially for patients with severe lung disease, revaccination of adults has not been recommended for routine practice. Pertussis begins with a runny nose, watery eyes, and cough, looking like an ordinary "head cold." It progresses, however, to a phase of severe paroxysmal coughing that can last for 8-12 weeks or longer.

The diagnosis is suggested by the severe paroxysmal nature of the coughing and often by a loud "whoop" heard on the post-tussive inspiration. If obtained within two weeks of the onset of cough, nasal culture, using special nasal swab and culture medium, may grow the bacteria. After two weeks, acute and convalescent serologies are needed to confirm the diagnosis. Treatment with macrolide antibiotics (e.g., erythromycin) reduces infectivity but has no effect on the cough. The cough is best treated with cough suppressants and possibly inhaled steroids. The value in making a specific diagnosis of pertussis is mainly in excluding alternative diagnoses and eliminating the need to test for them. Also, prophylactic treatment with macrolide antibiotics is recommended for persons with serious lung disease who have been exposed to a known case of pertussis.

Rarer Causes of Cough

Although the common causes of cough predominate, it is good to keep an open mind about rarer and potentially dangerous causes of cough in persons with a normal chest X-ray. These include lung cancer, interstitial lung disease, pulmonary hypertension, pulmonary embolism, thyroiditis, and foreign body aspiration (especially in children and the mentally retarded). Finally, material impacted in the auditory canal can trigger cough by irritating sensory receptors of Arnold's nerve, a branch of the vagus nerve.

Cough Suppressants

Cough suppressants can be divided into two types: those that act centrally at the level of the cough center in the brainstem and those that act on peripheral neural pathways. The two most commonly used centrally-acting agents are codeine (and its derivatives) and dextromethorphan. Clinical trials comparing the two compounds in awake subjects found no differences in efficacy. Codeine and its derivatives may be more effective for overnight cough suppression because of their greater sedative effect. For patients with persistent cough who are already taking narcotic analgesics (e.g., for pain control in metastatic cancer), addition of dextromethorphan may prove helpful because its effect is mediated through non-opioid receptors in the brain. Inhaled morphine has been tried for cough suppression but found to be only minimally effective. There are no opioid receptors in the lungs of humans, so that the observed cough suppression is probably due to systemic absorption of morphine after inhalation.

Peripherally-acting cough suppressants include inhaled lidocaine and oral benzonatate (Tessalon perles®). Inhaled lidocaine can be given by nebulizer (2.5 cc of a 4% solution) up to four times daily. In my experience, results with inhaled lidocaine have been mixed. Finally, expectorants (such as guaifenesin) and mucolytics (such as acetylcysteine or rh-DNase) are not indicated for the type of patient with chronic, non-productive cough that we have been discussing.

Case Example (concluded)

The patient had continued coughing and ultimately underwent repeat gastrointestinal evaluation. She was found to have delayed gastric emptying and persistent gastroesophageal reflux despite her initial surgery. After much debate, it was elected to proceed to repeat surgery, at which time she underwent a pyloroplasty and repeat fundoplication with tightening of the lower esophageal sphincter. Remarkably, she has been entirely cough-free ever since.


1. Irwin RS, Curley FJ, French CL. Chronic cough. The spectrum and frequency of causes, key components of the diagnostic evaluation, and outcome of specific therapy. Am Rev Respir Dis 1990; 141:640-7.

2. Poe RH, Israel RH, Utell MJ, Hall WJ. Chronic cough: bronchoscopy or pulmonary functio n testing? Am Rev Respir Dis 1982; 126:160-2.

About the author: Dr. Barbara Cockrill is a member of the Pulmonary and Critical Care Division at the Massachusetts General Hospital and Clinical Co-Director of the Partners Asthma Center there. She is an Instructor in Medicine at Harvard Medical School.

Return to Body of Text