Statistics from Altmetric.com
Factors associated with patient reports of reflux events
bile salts/pepsin (more relevant for mucosal injury)
liquid/gas (interacts with volume and distribution)
Oesophageal sensitivity increases distal to proximal (laryngo-pharyngeal structures very sensitive)
Increased in ENRD and functional heartburn
Decreased in Barrett’s CLO
Previous acid exposure
Everybody experiences gastro-oesophageal reflux on occasion. In health, reflux of air (“belching”) occurs most commonly during “transient lower oesophageal sphincter relaxations” (TLOSRs) triggered by gastric distension. Acid secretions and semi-digested food may also pass into the oesophagus during such events. Gastro-oesophageal reflux disease (GORD) is present when this reflux of gastric contents causes symptoms or mucosal damage.1 GORD patients do not necessarily have more TLOSRs than healthy controls.2 Rather, structural degradation and instability of the gastro-oesophageal junction increase the likelihood of reflux during TLOSRs and at other times (e.g. on straining).3 4 It is likely that the same changes allow greater volumes of gastric contents to pass the reflux barrier and to extend further into the oesophagus.5 6 Once reflux has occurred, ineffective motility and clearance are also important because prolonged exposure to acid and other noxious substances in refluxate (e.g. bile salts, pepsin) increase the risk of erosive reflux disease (ERD), Barrett’s columnar lined oesophagus (CLO) and other complications.7 8 Whether reflux triggers patient symptoms depends on a dynamic interaction between several factors, including patient age and sex, dietary factors, the volume, composition and distribution of the refluxate, mucosal disease, visceral sensitivity, and central factors including stress and patient vigilance (see box).5–20
Oesophageal pH testing was popularised by Johnson and DeMeester in the belief that GORD was not a symptom driven condition, but should be diagnosed by measurement of objective pathology.21 On this basis, the condition was conceived as a continuous spectrum of disease because oesophageal acid exposure is associated with the severity of symptoms and oesophagitis.22 Nevertheless, the relationship between reflux events, acid exposure, endoscopic findings and symptoms is not straightforward.13 For example, patients with Barrett’s CLO often have high levels of acid exposure but few symptoms because the metaplastic, columnar lining of the oesophagus is relatively insensitive to acid.23 Conversely, severe symptoms and a relatively poor response to acid suppression are often reported by patients with normal or near-normal acid exposure and without mucosal injury on conventional endoscopy.24 25
Ambulatory 24 h pH monitoring remains the standard investigation of GORD;26 however, it seems obvious that the value of pH studies is limited in patients in whom symptoms are not due to acid reflux. Multi-channel intra-luminal impedance (MII) detects and follows the movement of oesophageal contents and can distinguish fluid and gas within the lumen.27 Combined pH and impedance monitoring has shown that proton pump inhibitors (PPIs) reduce acid reflux but have no effect on the number of reflux events per se.28 Moreover, clinical studies have found that weakly acid or non-acid reflux is a common cause of persistent symptoms in patients on PPI treatment,29 including those with atypical symptoms and chronic cough.30 31 These findings led to rapid adoption of this technique in clinical practice; however, the analysis of MII data is time consuming and its place in the routine investigation of GORD is still being defined.
In this edition of Gut, Emerenziani and colleagues (see page 10.1136/gut.2007.130104) compare the findings of combined pH and impedance monitoring in patients with “endoscopy negative reflux disease” (ENRD), to those with ERD and healthy controls.14 Overall, and consistent with previous studies,29 32 about 80% of patient reports of heartburn and regurgitation were related to acid reflux. Moreover, the frequency of acid reflux as a proportion of all the reflux events was twice as high in patients than controls, adding to the evidence that increased gastric acid production, abnormal distribution of secretions or delayed gastric emptying play an important role in GORD.7 33 Oesophageal acid exposure was lower in ENRD than ERD; in contrast, the percentage association of reflux events with symptoms (symptom index) was higher for both acid and weakly acid reflux. On detailed analysis it was found that only a small proportion of symptoms (12%) were triggered by weakly acid reflux in ERD patients; however, this was significantly higher in ENRD patients (22%), especially in the subgroup with physiological oesophageal acid exposure (32%). In addition, and independent of acid content, the presence of gas in the refluxate (i.e. mixed reflux) increased the likelihood that symptoms were reported in ENRD patients, almost certainly because of increased refluxate volume and oesophageal distension. The importance of these findings is not to suggest that the acquisition of more and more complex information about oesophageal contents improves the diagnostic yield of GORD. Rather it is to emphasise that patients with a sensitive oesophagus can experience typical reflux symptoms in response to chemical or mechanical stimulation in the clinical setting, and that a high symptom index is a surrogate marker for visceral hypersensitivity and/or abnormal central processing of visceral sensations. As explained below, this insight may be of value in interpreting reflux studies and predicting the outcome of treatment.
Before recommending that combined pH and impedance replace conventional pH monitoring, technical factors that affect published comparison should be considered. Firstly, the use of antimony pH electrodes rather than “reference standard” glass electrodes in most studies (and most commercial catheters) could bias results. In clinical practice the diagnostic agreement between these systems is acceptable;34 35 however, pH measurements acquired by antimony electrodes (with external reference) drift upwards over time due to oxidation in acid environs.36–38 As a consequence, antimony electrodes may register less “acid reflux” events than glass electrodes, especially when gastric pH is elevated in the post-prandial period and on PPI treatment. Secondly, the recording characteristics and signal processing of pH and MII systems are fundamentally different. The former provides a continuous assessment of oesophageal acid exposure. Each pH measurement represents the mean acid exposure over a period of time (typically 6 s), with reflux events recorded after two consecutive readings under pH 4 (i.e. 12 s). The latter detects discrete acid and non-acid reflux events, but does not provide an assessment of “refluxate exposure” because current MII techniques are insensitive to volume change.39 In addition, impedance measurements are acquired at 50 Hz and reflux events are identified by a characteristic distal to proximal impedance fall (liquid) or rise (gas).40 Semi-automatic MII analysis software identifies reflux events, typically by an impedance fall >50% from baseline to <1000 Ω; however, findings must be checked manually and the baseline is routinely adjusted to increase sensitivity. A similar approach to pH data would allow rapid falls in pH that do not reach the pH 4 threshold to be counted as reflux events, an approach that identifies many symptom-associated reflux events currently detected “only” by MII, especially weakly acid reflux in patients on PPI treatment.52 Studies are needed that compare a similarly rigorous analysis of pH and MII data.
A further limitation of many studies is that reflux symptoms are not assessed independently of the pH and impedance measurements by validated questionnaire. The relationship between oesophageal acid exposure, symptom–reflux association and overall symptom severity is not clear. Patients with ERD or Barrett’s CLO and severe, prolonged oesophageal acid exposure often describe long periods of burning chest and epigastric discomfort, information that cannot be easily recorded by data loggers (especially at night), resulting in a low symptom index. In contrast, patients with ENRD and mild acid exposure more often experience discrete symptoms with reflux events, resulting in a high symptom index. This issue and the high day-to-day variability of symptoms probably explain why symptom–reflux association tests do not reliably predict the response to PPI treatment in unselected patients.28 An alternative approach is to consider the severity of oesophageal acid exposure and association of reflux events and symptoms (e.g. symptom index) as independent factors affecting overall symptom severity. Accordingly, pH monitoring provides a direct assessment of disease severity in terms of acid exposure, whereas symptom index (assessed by pH and/or MII monitoring) provides an assessment of visceral sensitivity to reflux events (fig 1). This approach is supported by a recent analysis of reflux and symptom events during 96 h wireless pH monitoring.41 As expected, symptom severity off treatment increased with oesophageal acid exposure (fig 2A). Almost all patients improved to some extent on high dose PPI treatment; however, the response was inversely correlated to symptom index (fig 2B). Thus, independent of oesophageal acid exposure, patients with a high symptom index (i.e. visceral hypersensitivity) were more likely to experience persistent symptoms on treatment due to weak acid and/or mechanical distension by persistent reflux episodes.41 These observations explain the paradox that patients with severe oesophageal acid exposure and mucosal disease often have a low symptom index but respond well to acid suppression, whereas patients with ENRD and functional heartburn with a high symptom index often continue to experience symptoms on treatment. Classification of patients on this basis is consistent with the shift in conceptual framework from GORD as a continuous spectrum of disease, to ENRD, ERD and Barrett’s CLO as different pathophysiological responses to acid exposure.1 15 At the same time it allows for some movement between ENRD and ERD observed in large longitudinal trials,20 42 as oesophageal acid exposure and visceral sensitivity vary over time with age, weight, stress and other factors.
In summary, research with combined pH and MII monitoring is slowly “making sense” of oesophageal contents and the occurrence of reflux symptoms; however, in clinical practice it is not clear whether this technique should be used routinely or reserved for special indications. Current guidelines advise a “treat and test” approach with further investigation by endoscopy and physiological measurement reserved for patients that fail to respond to acid suppression.43 A small number (∼2%) of patients (higher in Asian populations) have persistent acid reflux on treatment due to mutations in the cytochrome enzymes that metabolise PPIs.44 More commonly, persistent symptoms are reported by patients with hiatus hernia due to regurgitation of large volumes of gastric content and those with ENRD and functional heartburn due to hypersensitivity to ongoing reflux. Physiological studies are rarely needed to confirm the diagnosis in the former group unless fundoplication surgery is under consideration. In contrast, patients with no endoscopic findings often require investigation as “atypical” symptoms are common in these individuals and, conversely, reflux is not necessarily the cause of “typical” symptoms.29 The combined pH and MII technique is ideal for this indication; however, detailed assessment of the refluxate as described by Emerenziani and colleagues is not usually required as symptoms are not specific for oesophageal contents.14 Alternatively, applying the approach described above, pH studies alone may be adequate to detect oesophageal hypersensitivity if the recording is examined in detail to detect weakly acid reflux (pH 4–5), thus increasing the sensitivity of conventional symptom–reflux assessment. Indeed prolonged pH monitoring by wireless pH measurement may be preferable, especially in patients with intermittent symptoms.45 A study directly comparing the diagnostic yield of 24 h combined pH–impedance and 48 h wireless pH measurement is currently in progress.
Ultimately, the goal of investigation is to improve patient care. Previously, GORD treatment was directed towards healing mucosal injury. More recently, there has been a shift to controlling symptoms. Patients with ENRD with an incomplete response to PPIs have a form of functional bowel disease and, in common with these conditions, effective treatment can be directed not only at reducing the stimulus by acid suppression, but also at reducing visceral hypersensitivity. For those with ERD the aim will be “complete remission” of symptoms and mucosal disease, whereas in Barrett’s CLO the priority is to prevent progression to dysplasia and malignancy. This change of focus and the improved ability to identify the causes of symptoms and disease using modern physiological measurement will benefit patients by directing effective treatment on an individual basis.
The authors are grateful for the support of the Zurich Centre for Integrative Human Physiology.
Competing interests: None.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.