Dr M. F. Vaezi, Division of Gastroenterology and Hepatology, Center for Swallowing and Esophageal Disorders, Vanderbilt University Medical Center, C2104-MCN Nashville, TN, USA. E-mail: email@example.com
Gastro-oesophageal reflux disease is associated with several extraoesophageal disease states including laryngitis, asthma, chronic cough and non-cardiac chest pain. Currently, the exact role reflux of gastric contents play in the pathogenesis of extraoesophageal symptoms remain controversial.
Twenty-four hours pH monitoring is often considered the ‘gold standard’ in the diagnosis of gastro-oesophageal reflux disease and is increasingly utilized in patients with extraoesophageal symptoms. The use of this test is aimed at improving the association between patients’ extraoesophageal symptoms and oesophageal or hypopharyngeal acid reflux events. However, the clinical utility of pH monitoring in this patient population remains controversial.
Important clinical questions in this area include: does the presence of abnormal oesophageal acid reflux suggest a causal association between patients’ extraoesophageal symptoms and gastro-oesophageal reflux disease?
Conversely, does the absence of abnormal acid exposure in the oesophagus suggest lack of such an association? Should the test be performed on or off therapy and does it matter? In this study, the role of pH monitoring in laryngitis, asthma, chronic cough and non-cardiac chest pain is examined and answers to the above questions are addressed based on current data.
Establishing gastro-oesophageal reflux disease (GERD) as the cause of chronic laryngitis has been challenging to both gastroenterologists and ear, nose and throat (ENT) doctors. Patients with predominantly laryngeal symptoms such as chronic cough, sore throat, hoarseness, globus and excessive throat clearing are often diagnosed with GERD after laryngoscopy.1 However, a laryngeal examination may not be a specific marker for GERD diagnosis. Twenty-four hours pH monitoring is often the next test employed in this patient group to identify those with abnormal proximal or hypopharyngeal reflux events. Ohman et al.2 were among the first group to use ambulatory pH monitoring, in patients with contact ulcers in the larynx, identifying abnormal reflux events in 51% of patients. Twenty-four hours pH monitoring has been utilized in many studies of patients with laryngeal signs and symptoms (Table 1).3–16 However, whether or not this test plays a significant role in the diagnosis and ultimate management of this group of patients remains controversial.1
Table 1. Prevalence of laryngopharyngeal reflux by oesophageal pH monitoring
The overall pretherapy prevalence of abnormal pH findings in 15 studies of 1182 patients is reported to be 53% (Table 1). The prevalence of distal, proximal and hypopharyngeal pH monitoring are 42%, 44%, and 38%, respectively. The variability in prevalence of pH abnormalities reported in the literature may be the result of the differences in patient populations and the non-standardized placement of the pH probes. While some authors use manometry to locate the lower oesophageal sphincter (LES) and upper oesophageal sphincter (UES) for probe placement; others use direct laryngoscopy.1 The current recommendations suggest placing the hypopharyngeal probe 1–2 cm above the manometrically determined UES.1 The distal and proximal oesophageal pH probes should be placed 5 and 15 cm above the manometric LES.
Initial studies in patients with suspected GERD-related laryngitis focused on the role of proximal oesophageal pH probe. Jacob et al.17 studied the prevalence of abnormal proximal oesophageal acid exposure in 15 patients with typical GERD (group I); 15 patients with laryngeal symptoms but no physical findings on laryngoscopy (group II); and 10 patients with both laryngeal symptoms and laryngeal findings (group III). They demonstrated increased proximal oesophageal acid exposure in the subset of patients in both groups I and II, suggesting proximal oesophageal acid exposure may distinguish laryngitis patients from the typical GERD group.
As a direct measure of laryngeal extension of the reflux events, some investigators have turned to measuring the hypopharyngeal pH exposure. Koufman et al.4 found that seven of the 16 (44%) patients with laryngeal symptoms had an abnormal hypopharyngeal reflux event predominately in the upright position. Similarly, later studies by Shaker et al.18 and Ylitalo et al.,19 using hypopharyngeal pH probes, showed increased prevalence of pharyngeal reflux (PR) in patients with posterior laryngitis compared with control subjects and GERD patients (Table 1). Additionally, a recent study suggests that hypopharyngeal pH monitoring may be useful when used in combination with laryngoscopic findings to identify patients whose symptoms may be GERD-related. Oelschlager et al.20 evaluated 76 patients with respiratory symptoms suspected to be GERD-related. Patients were divided into three groups based on laryngeal reflux finding score (RFS) and PR events: group I – RFS−/PR−; group II – RFS+/PR− or RFS−/PR+; group III – RFS+/PR+. Improvement in symptoms after antireflux treatment occurred in 83% of patients in group III compared with only 44% of patients in group I.20 This finding suggests that GERD may be the likely cause of patients’ symptoms when both pH monitoring and laryngoscopy are abnormal. However, when both examinations are normal, GERD is most likely not playing a role in patients’ extraoesophageal symptoms. Therefore, it is possible that when hypopharyngeal pH monitoring is used with another mode of evaluation, its diagnostic utility may be increased. However, this needs to be tested in future studies.
While pretherapy prevalent studies suggest abnormal reflux events may be present in up to 53% of patients with reflux laryngitis (Table 1), it does not establish causal relationship. This may be due to poor sensitivity of pH monitoring. Vaezi et al.10 tested reproducibility and reliability of the proximal and distal oesophageal pH probe in 32 patients (11 controls, 10 with distal reflux and 11 with both proximal and distal reflux). They found that the sensitivity of distal and proximal pH probes were 70% and 55%, respectively. Additionally, a more recent study by Shaker et al.21 showed the number and duration of hypopharyngeal reflux events to be similar between the control subjects and patients with reflux laryngitis and vasomotor rhinitis. Some reflux into the hypopharynx may be a normal phenomenon. In a recent study by Bilgen et al.,22 there was no significant difference between hypopharyngeal acid exposure between 36 patients with laryngeal signs and symptoms and 23 healthy individuals. Additionally, Maldonado et al.23 found 10% prevalence of abnormal pH in normal subjects. Given this uncertainty, it is important to standardize oesophageal reflux parameters in the hypopharynx in order to better understand the causal relationship between reflux and laryngeal disease.
Importantly, abnormal findings on pH monitoring do not necessarily predict response to therapy. Ulualp et al.15 utilized hypopharyngeal pH monitoring in addressing the outcomes of acid-suppressive therapy in patients with posterior laryngitis with and without documented pharyngeal acid reflux events. In this study, using triple-probe pH monitoring, 28 of 39 patients were found to have PR events. However, both patients with and without documented pharyngeal acid reflux had improvement in their symptoms as well as in laryngeal findings with acid-suppressive treatment. Thus, the presence of hypopharyngeal reflux events prior to therapy did not predict treatment response. This was further supported by a recent placebo-controlled study of 145 patients treated with esomeprazole or placebo, finding that the degree of symptomatic or laryngeal improvement was independent of pretherapy pH monitoring in the oesophagus or the hypopharynx.24
The dichotomy in the literature regarding the clinical usefulness of hypopharyngeal pH monitoring may be due to several factors. (i) Probe position is highly operator-dependent and variable (direct visualization by laryngoscopy vs. measurement by manometry). (ii) Lack of consensus regarding duration and amount of reflux to denote abnormal PR. Some authors report any PR is abnormal, while others report small amounts of PR in healthy individuals.16 (iii) Poor sensitivity of pH monitoring in detecting acid reflux. This may be due to day-to-day variability of test and/or its less than adequate reliability as well as the possible intermittent nature of the reflux events (not recorded in only a 24-h period).15 Given these limitations, 24-h pH monitoring does not entirely establish or fully nullify reflux as the cause of laryngopharyngeal disease. Additionally, pH monitoring was not found to be clinically helpful in patients who continue to have symptoms while on proton-pump inhibitor (PPI) therapy.25 Among 115 patients who continued to have extraoesophageal symptoms while on b.d. therapy, the pretest likelihood of abnormal oesophageal acid exposure was reported to be no more than 2%. The value of other tests, such as prolonged wireless oesophageal pH probes and oesophageal and hypopharyngeal impedance monitoring, remains to be established in studies that are ongoing. Preliminary studies using impedance pH monitoring in patients with only extraoesophageal symptoms unresponsive to therapy did not find a role for non-acid reflux as the cause for patients’ continued symptoms and laryngeal abnormalities.26 However, this test was useful in 10–15% of patients who had a combination of typical and extraoesophageal symptoms. It has been shown that empiric treatment with a PPI proves to be a more reliable measure of the relationship between reflux and laryngopharyngeal symptoms. It is the improvement in symptoms, rather than the demonstration of acid reflux, that confirms this relationship.
Although there is little doubt regarding the association between asthma and GERD, the exact role of GERD remains undefined because its presence in many patients with asthma may not be causal. Nonetheless, multiple studies have shown that gastro-oesophageal reflux is common in asthmatics (Table 2).27–34 This may be due to either induced GERD during asthma attacks produced by decreasing intrathoracic pressure (i.e. asthma causing GERD) or GERD may contribute to the exacerbation of respiratory symptoms (i.e. GERD causing asthma; Figure 1). As an example of the latter, a study by Mansfield and Stein35 in 15 asthmatic patients undergoing intraoesophageal acid provocation testing demonstrated airway constriction on pulmonary function testing with reproducibility of patients’ reflux symptoms. Additionally, treating reflux symptoms with acid-suppressive therapy may also reduce asthma symptoms, suggesting a causal association. Goodall et al.36 illustrated significant improvement in reflux and night-time asthma symptoms in 14 of 18 patients with cimetidine treatment for 6 weeks. Similarly, Harper et al.37 demonstrated symptom improvement after cimetidine therapy in a group of 15 patients with both GERD and asthma.
Table 2. Prevalence of asthma-related reflux by oesophageal pH monitoring
There are two theories involving GERD leading to asthma. First, it is postulated that there is a ‘reflux mechanism’ promoting airway resistance and bronchospasm due to aspiration of gastric contents. Another theory suggests a ‘reflex mechanism’ by which acid reflux stimulates sensory afferents in the distal oesophagus, which then trigger a vagally mediated response of bronchospasm.30 It was thought that oesophageal pH monitoring may help identify acid-related bronchospasm of either mechanism. For example, in support of ‘reflex mechanism’, Gastal30 identified abnormal distal reflux in 44% in contrast to abnormal proximal acid exposure in 24% of asthmatics. Distal oesophageal acid exposure times were longer than that of the proximal oesophagus. Oesophageal pH monitoring may also allow for direct correlation between patients’ respiratory symptoms and oesophageal acid exposure. Harding et al.27 demonstrated acid reflux measured by dual probe pH monitoring in 118 of 164 (72%) asthmatics with reflux symptoms, compared with 10 of 35 (29%) asthmatics without reflux symptoms. In addition, this study showed a direct correlation between respiratory symptoms and reflux, as 119 of 151 (79%) respiratory symptoms and 43 of 67 (64%) wheezing episodes were associated with oesophageal acid exposure.
Several authors have used oesophageal pH monitoring to assess treatment outcomes with both medical and surgical therapy. Kiljander et al.34 used dual probe pH monitoring to assess the effect of omeprazole on asthma patients. Of the 107 patients who underwent pH monitoring, 57 (53%) were found to have abnormal pH values. Although 35% of patients reported improvement in their pulmonary symptoms on therapy, there was no improvement in pulmonary function. Similarly, using dual probe pH monitoring, Meier et al.31 found abnormal oesophageal pH parameters in 15 of 30 (50%) patients with asthma and only four of those 15 patients were shown to have improvement in forced expiratory volume in 1 s (FEV1) after treatment with omeprazole for 6 weeks. Thus, patients often report improvement in their asthma symptoms but objective improvements in asthma parameters generally are less impressive.
Most studies show that abnormal pH parameters prior to therapy fail to predict treatment response. However, Harding et al.33 found pH monitoring, especially in the proximal oesophagus, to be a useful clinical indicator of treatment response. They found that proximal oesophageal reflux predicted asthma response with a 100% sensitivity, 44% specificity, 79% positive predictive value and 100% negative predictive value. In this study, 30 patients with both asthma and GERD were treated with omeprazole 20 mg with titration of the dose until acid suppression was documented by 24-h oesophageal pH monitoring. They found that 22 patients (73%) responded to therapy as shown in improvement in asthma symptoms or peak expiratory flow (PEF). More importantly, frequent regurgitation or increased proximal regurgitation proved to be an excellent predictor of patients who responded to therapy.
Direct tracheal pH monitoring has also been evaluated in this group of patients to provide evidence for the aspiration theory of GERD asthma. Donnelly et al.29 studied three patients with severe asthma and GERD and found that decrease in tracheal pH correlated with a decrease in oesophageal pH. After antireflux surgery, asthma symptoms improved in two of three (67%) patients, and post-therapy decreases in oesophageal pH no longer lowered tracheal pH. In a similar study, Jack et al.32 evaluated both tracheal and oesophageal pH in four patients with asthma and GERD, where 37 falls in oesophageal pH were recorded. Only five of 37 (13.5%) of episodes of oesophageal acid exposure were followed by a fall in tracheal pH. Thus, pH monitoring in the trachea has not convincingly helped in better understanding the physiological relationship between GERD and asthma.
Therefore, studies in asthmatics have shown increased prevalence of GERD, based both on patient symptoms and on pH monitoring. However, the cause and effect relationship between asthma and GERD has been difficult to establish. GERD is believed to exacerbate asthma in a subset of patients; however, most studies suggest that pH monitoring prior to therapy does not predict response to treatment. Oesophageal pH monitoring may be considered in those who do not respond to aggressive medical therapy to rule out continued acid reflux. However, unless patients are non-compliant with taking their PPIs, pH monitoring on therapy in this group will most likely be normal.25 In some, impedance monitoring may be useful to determine if non-acid regurgitation may be responsible for patients’ symptoms. Recent studies suggest that in this group of patients, persistent regurgitation may be a good clinical marker for non-acid reflux.26 However, future studies in this area are clearly needed.
Chronic cough is one of the most common complaints presented to the primary care doctor.38 In those in whom medications, smoking and lung malignancy are ruled out, studies suggest five common potential causes: postnasal drip, asthma, GERD, chronic bronchitis and bronchiectasis.39 Moreover, Irwin and Madison40 found that chronic cough can be the sole presenting manifestation of GERD in up to 75% of patients. However, the association between cough and GERD remains controversial, as the exact pathogenesis is still unclear. Furthermore, the prevalence of GERD-associated cough has been variable and reportedly between 10% and 40%.41
Several studies suggest a significant relationship between chronic cough and GERD, based on temporal correlation between cough episodes and reflux events on pH monitoring (Table 3).42–48 Abnormal distal oesophageal acid exposure is reported with a prevalence of 92% in this group of patients. Early studies by Ing et al.42 using distal oesophageal pH monitoring in 13 patients with chronic cough reported higher reflux events in this group compared with controls. In addition, cough was found to be temporally related to more than 48% of the reflux episodes, while reflux was found to be temporally related to more than 90% of the cough episodes. Many of these patients did not report typical symptoms of GERD, such as heartburn and regurgitation, lending support to the theory of ‘silent GERD’. Paterson and Murat46 utilized hypopharyngeal and distal oesophageal pH probes and also suggested temporal relation between cough episodes and reflux events. Although only 9% and 1% of cough episodes were preceded by oesophageal and hypopharyngeal reflux episodes, respectively, coughing episodes were more frequently related to these reflux events. Similarly, Irwin et al.49 identified GERD as the sole presenting manifestation of chronic cough in adults 10% of the time.
Table 3. Prevalence of reflux-associated chronic cough by oesophageal pH monitoring
Oesophageal pH monitoring has been used in several studies to provide evidence for or against the aspiration and indirect (tracheobronchial cough reflex) theories.41 In support of the latter, studies by Irwin et al.44 documented significantly more GERD-induced coughs from the distal than the proximal oesophagus in a study group of 12 subjects. Additionally, three subjects who had coughs induced by distal reflux failed to demonstrate any proximal reflux. In addition, the investigators compared oesophageal pH monitoring with other tests such as endoscopy, GERD symptoms, barium swallow, chest X-ray and bronchoscopy and found that oesophageal pH monitoring was the test most frequently abnormal (sensitivity of 92%) and consistent with GERD-associated cough.44
However, some have questioned the utility of pH monitoring in patients with chronic cough. Using dual probe oesophageal pH monitoring, Waring et al.50 studied 25 patients with typical heartburn symptoms in addition to significant cough. They found that proximal oesophageal reflux was more common in those without atypical symptoms, suggesting poor predictive value for chronic cough. Furthermore, they found pH monitoring to be a poor marker of identifying response to therapy. Novitsky et al.51 further confirmed these results in 21 patients undergoing Nissen fundoplication using dual probe oesophageal pH monitoring. Only 36.2% and 23.4% of all cough episodes were induced by reflux events in the distal and proximal oesophagus, respectively and there was no difference in pretherapy pH parameter between responders and non-responders. Additionally, a more recent study by Ours48 in 17 patients with chronic cough and reflux by pH testing, only six (26%) patients responded to omeprazole therapy, finding a low sensitivity (35%) for pH monitoring in this group of patients. Furthermore, they showed that only 36% and 7% of cough episodes correlated with acid reflux in the distal and proximal oesophagus, respectively.
The role of pH monitoring post-therapy in patients initially unresponsive to therapy was highlighted by Poe and Kallay52 Fifty-six patients with GERD-associated cough were studied and the 12 non-responders to medical therapy were then evaluated by oesophageal pH monitoring. Of the 12 non-responders, 10 patients were found to have cough episodes related to reflux events. All but two of the non-responders were effectively treated with either prolonged or alternative medical therapy or surgery. Irwin et al.44 suggested chronic cough may be due to non-acid reflux events, and traditional pH monitoring may be normal in some patients. Thus, future studies with combined impedance and pH monitoring are needed in this group of patients. Until then, given the limitations of pH monitoring, patients with chronic cough should be treated empirically with PPIs for at least 2–3 months. In the unresponsive to therapy patients, pH monitoring on therapy may assess adequacy of medical treatment and potentially target those who may need altered therapy, including antireflux surgery.
Non-cardiac chest pain
Up to 30% of patients with chest pain have normal coronary arteries on coronary angiogram53 and GERD may be present in up to 60% of these patients.53–55 Oesophageal dysmotility may also be responsible for patients’ symptoms in a smaller subgroup.56 In the past, provocative tests such as acid perfusion or edrophonium testing have been used to establish symptom association with oesophageal acid exposure or motility disorders. However, these tests have limited roles due to their low sensitivity.53
Oesophageal pH monitoring and manometry have increasingly been used in patients with non-cardiac chest pain (NCCP). The overall prevalence of reflux by pH monitoring in this group is reported to be 41% (Table 4).54, 55, 57–63 Initial studies by Janssens et al.54 demonstrated that the combination of 24-h oesophageal pH and pressure readings increased the diagnostic yield (48%) when compared with manometry and acid perfusion tests (27%). Further studies by Hewson et al.55 found 24-h oesophageal pH monitoring to be the single best test for evaluating patients with NCCP. In this study, a positive symptom index (percentage of chest pain episodes associated with reflux) on pH monitoring identified GERD as contributing to chest pain in up to 50% of the patients. This was in sharp contrast to the diagnostic yield of the following tests: positive acid perfusion test (10%), positive edrophonium (10%), both tests combined (5%) and manometry and provocation tests (28%). Of the 100 patients in this study, 24-h pH testing provided a diagnosis in 31 patients in whom other diagnostic tests such as acid perfusion, edrophonium testing and manometry were inconclusive.55
Table 4. Prevalence of reflux in non-cardiac chest pain (NCCP) patients by oesophageal pH monitoring
More recently, some studies have suggested that pH monitoring has low diagnostic yield in patients with NCCP. Soffer et al.56 found that only 20.7% of all chest pain episodes in his patients were related to reflux events. Additionally, only 27.1% of chest pain episodes were associated with either reflux or abnormal oesophageal motor events. Lacima et al.61 confirmed these findings in a recent study where GERD was reported to be the cause in 12% of 90 patients with NCCP. Although they found abnormal pH parameters in 43% of patients, a direct association was found in only a minority of the patients. Furthermore, only 25 of 144 (17%) of chest pain events correlated with oesophageal exposure to pH <4.0. Breumelhof et al.57 investigated a group of 44 patients with NCCP and found 33% of the 111 total chest pain episodes to be related to reflux and 23.4% to abnormal motility. More importantly, using the symptom index of ≥75%, only two patients were found to have pain related to reflux. In 36 patients (81.8%), no relation was identified between patients’ chest pain and either oesophageal reflux or motility disorders.
The role of pH monitoring and treatment outcome in patients with NCCP has been investigated in some studies. Lam58 studied a group of 41 patients initially admitted to the coronary care unit with suspected coronary artery disease whose pain was later shown to be non-cardiac in origin. Pain was related to reflux in 13 patients (43%) and motility disorders in 10 patients (33%). More importantly, this group found that 11 of 13 (85%) reflux-related NCCP patients responded to medical treatment and eight of 10 (80%) patients with motility abnormalities showed improvement of chest pain with spasmolytic therapy. Similarly, Fass et al.60 found that the greater the oesophageal acid exposure, the greater the likelihood of improvement in symptoms on omeprazole therapy. However, initial empiric therapy with PPIs is believed to be more cost-effective and may have similar sensitivity and specificity.53 In addition, while pH monitoring does demonstrate reflux in patients with NCCP, it does not establish causality, as would sustained improvement in symptoms on acid-suppressive therapy. Earlier studies by Achem et al.62 identified GERD using distal oesophageal pH monitoring in 13 of 20 patients with NCCP. Twelve of these patients underwent 8 weeks of medical therapy with either ranitidine or omeprazole. They showed that despite normalization of oesophageal acid reflux parameters confirmed by on-therapy distal oesophageal pH monitoring, two patients still remained symptomatic. In a later study by Achem et al.,63 34 patients were treated with an 8-week course of omeprazole with a resultant significant reduction in chest pain severity and fraction of chest pain days. However, more importantly, distal oesophageal pH monitoring was shown to have no predictive value on treatment outcome.
In summary, there is marked variability in the published studies by different authors regarding the prevalence of reflux in NCCP patients using 24-h pH monitoring. There are several factors accounting for these findings. First, each study group consists of patients with variable frequencies of chest pain episodes. For example, Lam58 found a higher prevalence of reflux in patients with NCCP as his patient population consisted of those admitted to the coronary care unit and would, therefore, have a higher frequency of chest pain. Breumelhof et al.57 recommended that 24-h pH monitoring be restricted to those patients with daily chest pain so as to increase the diagnostic value. The intermittent nature of reflux and chest pain episodes also precludes accurate association between the two entities over the relatively short period of 24 h. Additionally, not all reflux episodes are associated with chest pain. Despite the high prevalence of reflux in patients with NCCP, the actual correlation between chest pain episodes and reflux (symptom index) may be less impressive.64 Lastly, the role of 24-h pH monitoring as a clinical predictor of treatment outcome remains to be more clearly elucidated. Once again, 24-h pH monitoring proves to be of marginal utility prior to therapy. One can argue that it should be reserved for those patients with daily chest pain who are unresponsive to aggressive therapy.
GERD may be one cause of symptoms such as hoarseness, sore throat, cough, asthma and chest pain. Twenty-four hours pH monitoring may establish the presence of acid in the oesophagus or hypopharynx of such patients; however, its role in causally associating patients’ symptoms to GERD is still controversial. Most studies suggest that pretherapy pH abnormalities do not predict response to medical or surgical therapies. Thus, the current recommendations suggest aggressive acid suppression initially in all those suspected of having GERD and reserving pH monitoring to those unresponsive to therapy. The role of non-acid or intermittent acid reflux in this group is not well studied and deserves future investigation. Until this point in time, the surgical intervention should only be offered to those in whom the cause and effect relationship between GERD and patients’ symptoms is well documented.