Partners Asthma Center Grand Rounds
Richard S. Irwin, M.D.
Chronic Cough Due to
Gastroesophageal Reflux Disease
Roughly twenty years ago, when my colleagues and I reported on a group of patients with cough-variant asthma, we grew interested in the spectrum and relative frequency of various causes of chronic cough. We defined chronic cough as a persistently troublesome cough of at least three weeks’ duration in the ambulatory setting. We sought to develop a systematic protocol to evaluate patients with chronic cough, and we based our initial protocol on current concepts of the anatomy of the cough reflex (Figure 1). At the time we included the stomach as a site for receptor activation of the cough reflex because of the observation in the pediatric medical literature that children often cough immediately before vomiting.
In the late 1970s our evaluation of the gastrointestinal (GI) tract in patients with chronic cough included endoscopy, barium swallow, and upper GI series but no 24-hour esophageal pH probe monitoring. At that time equipment for ambulatory pH monitoring was not widely available. When we reported our experience with the causes of cough 20 years ago, we found that post-nasal drip, asthma, and chronic bronchitis from cigarette smoking were the most prevalent etiologies. In this series gastroesophageal reflux disease (GERD) was present in only 10% of cases, and these patients were easily diagnosed. They all had daily heartburn, sour taste in the mouth, and symptoms of regurgitation, and they all got better with simple anti-reflux treatments (type 2 histamine blockers, antacids, Gaviscon®, and an anti-reflux diet).
After this initial success of our anatomically-based diagnostic protocol, I began to see more difficult cough patients in referral. One particular patient had a dramatic influence on my thinking about chronic cough.
This 56 year-old woman had been treated without success for post-nasal drip syndrome secondary to what was thought to be perennial non-allergic rhinitis. She had had a non-productive cough for 2 ½ years. While she complained of a post-nasal drip, she denied throat clearing, allergies, wheeze, shortness of breath, or symptoms consistent with reflux disease or swallowing difficulties. She had never smoked, and her home and work environments could not be implicated in her cough. She was taking no medicines, specifically no angiotensin converting enzyme(ACE) inhibitors. Her physical examination was normal except for minimal cobblestoning of the posterior pharyngeal wall suggesting chronic irritation. Chest x-rays, sinus x-rays, methacholine challenge, and bronchoscopy were all normal. She had been treated for rhinitis with antihistamines, decongestants, and nasal steroid sprays, all without benefit.
Although it was clear that something was irritating her oropharynx, treatment for post-nasal drip had not helped her cough, and she had no GI symptoms to suggest irritants coming up from below. Nonetheless, I ordered a barium swallow and was surprised to find that she had reflux of gastric contents to her mouth. She had such severe esophageal reflux that it took her 15-20 swallows before she could get the first gulp of barium down and have it stay down. She was treated with ranitidine, metoclopramide, and a high protein/low fat anti-reflux diet, and her cough totally disappeared.
Cough due to Silent GERD
My experience with this patient stimulated my interest in studying further the relation between GERD and cough. I sought the help of my gastroenterologist colleague, Dr. John Zawicki, and together we developed a protocol for 24-hour esophageal pH probe monitoring of patients with chronic cough who failed to get better with application of our standard evaluation and treatment protocol. We were quite naïve about this technique of esophageal pH probe monitoring, and perhaps this inexperience allowed us to interpret the results in novel ways, making it possible for us to find evidence of esophageal reflux where other gastroenterologists had not.
Over 18 months we identified 34 patients with chronic cough that did not improve despite application of our standard protocol. On esophogeal pH probe monitoring, 27 of the 34 patients had very abnormal studies, indicating GERD. With medical anti-reflux therapy, their cough disappeared. Nine of these patients agreed to follow-up esophageal pH probe monitoring after resolution of their cough. Probes were placed in both the proximal and distal esophagus. There were 4 men and 5 women in the group, with a mean age of 52 years. They had been coughing on average for 33 months. None of them ever had any GI complaints. Three of them had had barium swallows that showed minimal reflux into the distal esophagus, considered within the normal range of findings. Four of the nine came to us with difficult-to-control, cough-variant asthma, including one patient taking prednisone 100 mg/day.
In these nine patients studied both before and after therapeutic invervention, we demonstrated that the number of refluxes in 24 hours decreased significantly and correlated with reduction in the frequency of reflux-induced coughs. Specifically, the amount of time that the pH went to less than 4 in the distal esophagus correlated with decreased cough frequency, suggesting the possibility of an esophageal-bronchial reflex. The average time until resolution of cough was 5 ½ months, consistent with other studies that also indicate very gradual improvement with treatment. We concluded from these observations that cough may be the sole or predominant manifestation of GERD – “cough-variant GERD” – and that 24-hour esophageal pH probe monitoring is a safe and effective method for its diagnosis.
When we again prospectively studied the spectrum of causes of cough and their relative frequencies, we incorporated 24-hour esophageal pH probe monitoring into our diagnostic protocol. Approximately 10 years after our initial series, we reported on 102 patients with chronic cough whom I personally saw over one year (Figure 2). As in most reported series regarding chronic cough, the number of women exceeded the number of men (59:43). In 26% of patients more than one etiology was found to account for the cough. We have encountered patients in whom cough did not resolve until 5 different contributing factors were addressed. As in our previous series, post-nasal drip was the most common cause of cough, and asthma was second. In this series GERD was the third most common cause of chronic cough, and among this subgroup with GERD, silent GERD or cough-variant GERD accounted for 43% of the cases. Among patients with no symptoms of esophageal reflux in whom barium esophagography is negative, 24-hour esophageal pH probe monitoring is an important test to establish the possibility that GERD is causing cough.
Three Additional Questions to be Answered
As we continued to explore the relation between GERD and cough, we sought to address three additional questions.
1) Is there a clinical profile that accurately predicts when cough is due to GERD? The idea of a clinical profile suggesting cough due to GERD derived from the observation that chronic cough was most often caused by post-nasal drip, asthma, or GERD. We reasoned that one should be able to predict cough due to GERD among patients who had the following features: they did not take an ACE inhibitor; were non-smokers; had a normal chest X-ray; had a negative methacholine challenge (or persistent cough despite appropriate treatment for asthma); and had a persistent cough after receiving treatment for post-nasal drip due to rhinosinus disease, including therapy for an infectious sinusitis, if necessary (Table 1).
2) How best should we interpret the findings on 24-hour esophageal pH probe monitoring when using the test to diagnose cough due to GERD? Most gastroenterologists view pH probe monitoring as a means of assessing whether acid reflux is causing esophagitis. Patients with cough due to reflux disease very infrequently have esophagitis; the findings on pH probe monitoring need to be analyzed differently. We consider a positive result – indicating that GERD is causing cough – to occur when the esophageal pH drops below 4 and when the patient then coughs within 3 minutes of the event.
3) What is the mechanism by which GERD causes cough? With respect to pathogenesis, GERD might cause cough when gastric contents reflux to the level of the hypopharynx and are then aspirated; it might trigger cough receptors in the larynx; it might stimulate an esophageal-bronchial reflex when gastric acid refluxes onto the distal esophageal mucosa, and it might cause cough by a mechanism not at all dependent on the acidity of refluxed contents. We sought to test these latter possibilities by performing a randomized, double-blind Bernstein test among our patients with chronic cough due to GERD.
To address these three issues, we recruited 12 patients with unexplained chronic cough, selected on the basis of the clinical profile described above that predicted GERD as the etiology of their cough. All patients received a careful history, physical examination, chest X-ray, four-view sinus films, barium swallow, 24-hour esophageal pH probe monitoring, upper GI endoscopy, esophageal manometry, and fiberoptic bronchoscopy. All 12 patients had cough due to GERD based on the results of these diagnostic tests, with confirmation established by resolution of their cough with treatment for GERD. Only 3 of the 12 patients had any GI complaints. Four patients had positive barium esophagrams, defined as the presence of unprovoked reflux to the mid-esophagus or higher.
Esophageal pH probe monitoring was positive in 11 of the 12 patients, indicating that the test is not 100% sensitive for diagnosing GERD-induced cough. The one exception was a patient with an entirely normal pH probe study (findings within the 95% confidence intervals for normal published values) whose barium swallow showed reflux to the thoracic inlet. From this patient it became clear that reflux into the esophagus will not be detected by pH probe monitoring unless the pH of the gastric contents is abnormal. Nonetheless, reflux in this patient caused her cough, raising the possibility that alkaline reflux may also provoke cough. This possibility remains debated in the medical literature.
Bronchoscopy showed edema of the vocal cords and proximal airways in only two patients, findings that might be attributed to repetitive coughing as much as to aspiration. In addition, most patients infrequently had reflux into the proximal esophagus and two patients never did. We further evaluated the pathogenetic role of acid in these patients with cough due to GERD by performing double-blind Bernstein tests. We recorded the number of coughs after normal saline or 0.1% hydrochloric acid was applied to the distal esophagus via nasal tube. We found no difference in cough frequency during the two different periods. Even among patients with GERD-induced cough, acid applied to the distal esophagus did not consistently provoke cough. Overall, our data are most consistent with the hypothesis that GERD causes cough by stimulation of esophageal mucosal receptors rather than via aspiration, and that this stimulation occurs predominantly in the distal rather than proximal esophagus. Furthermore, in the majority of cases, acid is unlikely to be the sole mediator or trigger of reflux-induced cough.
Surgery for GERD-Induced Cough
As I began to see more and more patients with cough due to GERD, I encountered a patient who fit our clinical profile for GERD-induced cough despite a borderline positive esophageal pH probe study. Her cough continued despite intensive anti-reflux medical therapy and diet. She expressed that “my life is being destroyed by my cough.” Given the lack of success of all of our medical interventions, we referred her for surgery. She underwent a laparoscopic Nissen fundoplication, and her cough disappeared. This experience led us to explore prospectively the one-year outcomes from anti-reflux surgery performed for chronic cough due to GERD that is resistant to intensive medical therapy.
Patients were recruited into this study if despite maximal medical therapy for at least three months they still had a positive 24-hour esophageal pH probe test or if their cough persisted to a degree that was considered to preclude a satisfactory quality of life. They had an extensive pre-operative GI evaluation, including flexible laryngoscopy and bronchoscopy, consultation with a gastroenterologist, barium esophagography and gastric emptying study, upper endoscopy, and esophageal manometry. Finally, we insisted on referral to a surgeon experienced in performing laparoscopic and open anti-reflux operations.
At this time we have experience with 16 patients who completed this protocol with one-year follow-up. Four of these patients had had prior laparoscopic surgery for GERD but continued to cough. Because they fit the clinical profile for cough due to GERD, they underwent further evaluation that led to their inclusion in this study. Three of these four patients had not had gastric emptying studies performed prior to their original operation. We found that they had very delayed gastric emptying, and a simple gastric drainage procedure successfully alleviated their GERD and their cough.
Of these 16 patients, 14 had a decrease or disappearance of their cough one year after surgery. One patient had no change, and one worsened. Overall, patients improved significantly on their sickness impact profile, both in terms of psychosocial and physical impairment. Improving cough severity and quality of life in the short term (2 months postoperatively) predicted sustained improvement at one year. We also learned from this study that cough may be due to GERD even among patients who do not improve with intensive anti-reflux medical therapy. At present, I consider intensive anti-reflux medical therapy to include high doses of a proton pump inhibitor plus a motility agent such as metoclopramide, and I also perform a repeat esophageal pH probe study to ensure that gastric acid has been fully suppressed.
In a retrospective analysis of patients with cough due to GERD who only had resolution of their cough after undergoing anti-reflux surgery, we found that some patients had complete suppression of gastric acid (and of events with esophageal pH less than 4) with intensive medical therapy, despite continued cough. Some of these patients had relief of their GI symptoms with intensive medical therapy, others did not. We have learned that even among patients who report total relief of their heartburn on medical therapy, their continued cough may be due to GERD. Since reflux of gastric contents other than acid is likely to cause cough in this setting, the term “acid reflux disease” is probably a misnomer when applied to at least some patients with reflux-induced cough.
To summarize, cough due to GERD can be suspected in patients with typical GI symptoms of reflux or with a typical clinical profile that raises suspicion of “silent GERD.” Further diagnostic testing may include 24-hour esophageal probe monitoring and barium esophagography. Empiric therapy is appropriate for patients fitting the typical clinical profile, but if cough does not resolve on intensive medical therapy, then further diagnostic testing (esophageal probe and barium esophagram) becomes necessary. Patients with persistent cough despite appropriate medical anti-reflux therapy may have inadequate suppression of gastric acid, may have continued reflux requiring surgical intervention, or may have additional causes of cough that require simultaneous treatment.
Intensive medical treatment includes a high protein/low fat anti-reflux diet, anti-reflux lifestyle changes (e.g., avoidance of large meals before recumbency and sleeping on an incline with the head elevated above the stomach), twice-daily therapy with a proton pump inhibitor (taken ½ hour before breakfast and before the evening meal), and treatment with a prokinetic agent. Treatment of co-morbid illnesses may be important, especially obstructive sleep apnea syndrome. Finally, one should avoid medications that decrease lower esophageal sphincter tone, including theophylline, nitrates, and calcium channel blockers.
Irwin RS, Madison JM. The diagnosis and treatment of cough. N Engl J Med 2000; 343:1715-21.
Irwin RS, Richter JE. Gastroesophageal reflux and chronic cough. Am J Gastroenterol 2000; 95 (8 Suppl): S9-14.
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.
About the Author: Dr. Richard Irwin is Director of the Pulmonary and Critical Care Division at the University of Massachusetts Medical School in Worcester. He is Professor of Medicine at the University of Massachusetts. He is recognized worldwide as a leading authority on the diagnosis and management of cough
Figure 2: Spectrum of Causes of Chronic Cough in a Prospective Series of 102 Patients (PND = postnasal drip syndrome; GER = gastroesophageal reflux)