Dr G. E. E. Boeckxstaens, Academic Medical Centre, Division of Gastroenterology and Hepatology, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands. E-mail: firstname.lastname@example.org
The oesophago-gastric junction functions as an anti-reflux barrier preventing increased exposure of the oesophageal mucosa to gastric contents. Failure of this anti-reflux barrier results in gastro-oesophageal reflux disease, and may lead to complications such as oesophagitis, Barrett’s oesophagus and eventually oesophageal carcinoma.
Recent studies have suggested that transient lower oesophageal sphincter relaxation is the main mechanism underlying gastro-oesophageal reflux. It involves a prolonged relaxation of the lower oesophageal sphincter, mediated by a vago-vagal neural pathway, synapsing in the brainstem.
Several drugs, such as atropine, baclofen and loxiglumide, have been shown to reduce the rate of transient lower oesophageal sphincter relaxations and concomitantly the number of reflux episodes. These findings illustrate that transient lower oesophageal sphincter relaxations may represent a potential new target for the pharmacological treatment of gastro-oesophageal reflux disease.
It is possible that the reduction in the number of transient lower oesophageal sphincter relaxations may also contribute to the beneficial effect of fundoplication and new endoscopic anti-reflux procedures. It should be emphasized, however, that other factors, such as low lower oesophageal sphincter pressure, the presence of a hiatal hernia and impaired oesophageal peristalsis, are also of great importance. Therefore, whether the targeting of transient lower oesophageal sphincter relaxations is the ‘golden bullet’ in anti-reflux therapy remains to be proven, as evidence of an effective control of gastro-oesophageal reflux in reflux patients is still lacking.
Gastro-oesophageal reflux disease (GERD) is a disorder characterized by an increased exposure of the oesophagus to the intragastric contents. Typical reflux symptoms, such as heartburn, regurgitation or retrosternal pain, are reported by 5–20% of the population at least once a week, illustrating that gastro-oesophageal reflux is a common disorder.1–3 The majority of patients have mild to moderately severe complaints. However, increased exposure of the oesophageal epithelium to the gastric contents may lead to complications, such as oesophagitis, peptic strictures, Barrett’s oesophagus and eventually oesophageal carcinoma.4–6 As GERD is a chronic disorder, the majority of patients (61%) will require acid suppressive maintenance therapy or, ultimately, surgery (fundoplication).7
DEFENCE MECHANISMS AGAINST GASTRO-OESOPHAGEAL REFLUX
The pressure gradient between the abdomen and the thorax creates a situation in which intragastric contents are continuously driven towards the oesophagus. Furthermore, this pressure gradient increases during inspiration and abdominal straining. Therefore, a pressure zone with the capability to compensate for elevations in intragastric pressure is necessary to prevent gastro-oesophageal reflux. This dynamic pressure zone is generated by the oesophago-gastric junction. Both the lower oesophageal sphincter (LOS) and the crural diaphragm are part of this sphincter complex and contribute to the defence against gastro-oesophageal reflux. The LOS, or internal sphincter, is a specialized part of the circular smooth muscle of the distal oesophagus, with a length of approximately 4 cm. In healthy volunteers, this part of the sphincter complex generates a tonic pressure of 15–30 mmHg above the intragastric pressure.8,9 As reflux mainly occurs when the LOS pressure is below 5 mmHg, this should be sufficient to protect against the pressure gradient between the stomach and the oesophagus.10 However, with increasing abdominal pressure, such as during inspiration and straining, an additional compensatory mechanism is required. This function is fulfilled by the crural diaphragm, which encloses the proximal 2 cm of the LOS and contracts during each inspiration. The crural diaphragm will not only prevent ‘stress’ reflux, but also reflux during periods of absent LOS pressure, indicating the important function of this external part of the sphincter complex in maintaining an effective anti-reflux barrier.11–12
Once the gastric contents have passed the oesophago-gastric junction, the exposure time of the oesophageal epithelium should be limited as it is not equipped to withstand prolonged exposure to the noxious mixture of gastric acid, pepsin and, occasionally, bile. During a reflux episode, one or two peristaltic waves will empty the distal oesophagus and only a fraction of the refluxate will remain. However, the pH is still low after the initial peristaltic action and the oesophageal pH will only be restored after frequent swallowing, due to the buffering capacity of saliva.14–16
First of all, it should be noted that gastro-oesophageal reflux itself is a physiological phenomenon occurring regularly in healthy subjects.17,18 Several studies have reported oesophageal damage and inflammation in relation to rising levels of oesophageal acidification, with most reflux occurring in patients with Barrett’s oesophagus.4,19,20 Therefore, acid reflux is considered to be pathological if the oesophagus is exposed to acid (or pH < 4) during more than 5.8% of the time during 24-h pH-metry.21 As the majority of GERD patients have a normal production of gastric acid, it is most likely that the above described defence mechanisms are ineffective in these patients.22
LOWER OESOPHAGEAL SPHINCTER
Originally, it was commonly believed that a defective LOS pressure was the main mechanism underlying reflux in GERD patients. However, it should be emphasized that LOS pressures in the normal range are also recorded in a substantial number of GERD patients. It was not until 1982 that Dodds and co-workers unravelled this discrepancy, and showed that the majority (65%) of GERD patients have reflux episodes during transient relaxation of the LOS, confirming their earlier observation in healthy volunteers.10,23 Transient lower oesophageal sphincter relaxations (TLOSRs) are manometrically defined as prolonged relaxations of the LOS which are not associated with a swallow24 (Figure 1). During a TLOSR, the activity of the crural diaphragm is also inhibited, presumably facilitating the occurrence of gastro-oesophageal reflux (see below).12 Distension of the stomach with free air, an intragastric balloon or a meal will result in an immediate increase in TLOSRs.25–27 By dividing the stomach into different compartments, Franzi et al. showed that the cardiac region of the stomach has the lowest threshold for the triggering of TLOSRs.28 Vagal cooling abolished the triggering of TLOSRs, suggesting that they are mediated via a vago-vagal reflex.29 On the other hand, splanchnicectomy did not influence the rate of TLOSRs, indicating that splanchnic nerves are not involved in mediating these relaxations.30 Based on these experiments, a neural pathway underlying TLOSRs has been suggested. The afferent pathway of the reflex arc consists of stretch-sensitive vagal afferent fibres presumably located within the muscle layers of the stomach.31 The vagal afferent fibres activated during gastric distension terminate in the nucleus tractus solitarius and dorsal motor nucleus of the vagal nerve in the brainstem.32–34 The neurones located in the nucleus tractus solitarius make synapses with dendrites from the motor neurones located in the dorsal motor nucleus of the vagal nerve, which in turn project to the enteric nervous system of the LOS35 (Figure 2).35–3637 This circuitry allows fast vago-vagal reflexes and is presumably involved in mediating TLOSRs.
As TLOSRs account for the majority of reflux episodes, GERD patients should have more TLOSRs. However, studies have reported conflicting results with both increased and normal TLOSR rates in GERD patients.25,38,39 Evidence for an increased percentage of TLOSRs accompanied by reflux appears to be more convincing.10,38,39 Why GERD patients have an increased percentage of TLOSRs accompanied by reflux remains unclear. Possibly, intragastric factors and/or abnormal crural diaphragm activity may be involved.
As already mentioned, the oesophago-gastric junction is composed of two closely related structures, the LOS and the crural diaphragm. In GERD patients, it is a common endoscopic and radiological finding that the sphincter complex is physically separated, i.e. hiatal hernia. In this situation, the stomach has migrated through the diaphragm into the thorax, thereby separating the LOS and crural diaphragm.40,41 Detailed manometric analysis of the hiatal hernia shows an altered topographic pressure profile with two distinct high pressure zones, indicating that the oesophago-gastric junction is also divided into two distinct pressure zones: the proximal zone representing the LOS and the distal zone representing the crural diaphragm.42 A simulated reduction of the hiatal hernia shows that reposition of the two individual high pressure zones leads to normalization of the oesophago-gastric junction pressure profile, indicating that both structures are intact, but separated. Furthermore, a hiatal hernia is associated with a decreased LOS pressure and gastric contents may accumulate in the hiatal sack, thereby facilitating reflux during swallow-induced LOS relaxation.43 In addition, hiatal hernia patients have impaired oesophageal peristalsis, thereby decreasing oesophageal clearance. 44
The link between the functional and the anatomical phenomena related to reflux, i.e. TLOSRs and hiatal hernia, was made only very recently. Kahrilas et al. showed that GERD patients with a hiatal hernia had significantly more TLOSRs compared to non-hiatal hernia patients. Furthermore, they showed a positive correlation between the distance between the squamocolumnar junction and hiatal centre and the rate of TLOSRs induced by gastric distension.45 Thus, in the presence of a large hernia, basal LOS pressure is absent, the compensatory function of the crural diaphragm is lost, clearance is impaired and the rate of TLOSRs is increased. The combination of these factors most likely explains the increased incidence of reflux in these patients.
The crural diaphragm plays an important role in the function of the anti-reflux barrier in an intact oesophago-gastric junction. Both animal and human studies show that pharmacological or surgical abolition of the LOS pressure alone does not induce gastro-oesophageal reflux.46,47 However, an additional crural myotomy results in a significant increase in reflux in rats. Similarly, reflux only occurs during absent LOS pressures (LOS pressure < 5 mmHg) if crural diaphragm activity is simultaneously inhibited, a phenomenon also observed during a TLOSR.47 In contrast to healthy subjects, however, little is known about the relation between the function of the crural diaphragm and the occurrence of reflux in GERD patients. Whether crural diaphragm activity is impaired in GERD patients is not known. Further experiments are therefore of eminent importance as the inhibition of crural diaphragm activity seems to be a major determinant for gastro-oesophageal reflux, at least when LOS pressure is absent.
The inhibition of the crural diaphragm activity during a TLOSR is most likely triggered via the vagal afferent pathway, which also mediates TLOSRs48 (Figure 2). The afferent fibres terminate in the brainstem and are thought to inhibit the respiratory nuclei. However, no immunohistochemical studies are yet available to provide anatomical evidence for this pathway. In contrast, a recent study even suggests the existence of a peripheral inhibitory pathway, as both phrenectomy and vagotomy did not prevent crural diaphragm inhibition during oesophageal distension.49–51
The stomach may play an important role in the pathogenesis of gastro-oesophageal reflux. GERD patients have prolonged gastric relaxation after meal ingestion.52 Stretch of the cardiac region is the major trigger for TLOSRs, and prolonged distension of the proximal stomach may therefore result in a higher number of post-prandial TLOSRs and reflux episodes.52 Impaired gastric emptying and prolonged storage of food in the fundic region are observed in GERD patients.53 It may be speculated that prolonged accumulation of food in the fundus may trigger more TLOSRs. Indirect evidence is provided by the observation that infusion of the 5-hydroxytryptamine1 agonist, sumatriptan, is associated with an increase in the rate of post-prandial TLOSRs and reflux episodes, most likely due to fundic relaxation and decreased gastric emptying. Alternatively, and probably more importantly, this might result in an increase in the percentage of TLOSRs accompanied by reflux, as suggested by the observation that placement of an intragastric balloon (500 mL) increased the percentage of reflux during TLOSRs from 31 to 100%, resulting in an increase in total acid exposure time.54
TARGETING TLOSRS AS A TREATMENT FOR GASTRO-OESOPHAGEAL REFLUX
In 1995, Mittal et al. showed for the first time that reducing TLOSRs in humans was feasible, resulting in a decrease in reflux.47 The reduction in TLOSRs and reflux during infusion of the muscarinic receptor antagonist atropine was later confirmed in GERD patients, indicating that reducing reflux by inhibiting TLOSRs might have therapeutic implications.55 A reduction in TLOSRs can be obtained by intervening with the afferent, central or efferent arm of the reflex arc. The latter will impair swallow-induced LOS relaxation, resulting in a situation comparable to achalasia, and is therefore not preferred. Reducing the trigger of TLOSRs, i.e. distension of the stomach, or intervening with the vagal afferent fibres and/or the motor pattern localized in the brainstem are hypothetically better options. Before discussing the studies evaluating intervention with the triggering of TLOSRs, it is important to emphasize that changes in gastric tone or gastric emptying may drastically affect the rate of TLOSRs. Therefore, as changes in TLOSRs may be secondary to changes in stimulus intensity, the monitoring of gastric tone or emptying is essential for a proper evaluation of the effects of such intervention studies.
A reduction in TLOSRs could theoretically be obtained by intervention with the underlying neurocircuitry as described above. Developing drugs targeting TLOSRs requires insight into the neurotransmitters involved in this neuronal pathway. Immunohistochemical studies have provided valuable information on the neurotransmitters and their receptors localized in the vagal afferent fibres, the nucleus tractus solitarius and the dorsal motor nucleus of the vagal nerve. At present, animal and human studies have provided evidence for a role of (at least) acetylcholine, cholecystokinin (CCK), γ-aminobutyric acid (GABA), glutamate, nitric oxide (NO) and opioids30,47,55–70 (Figure 3). As discussed above, atropine was the first drug which significantly reduced the rate of TLOSRs in both healthy volunteers and GERD patients.47,55 The investigators used atropine to artificially abolish LOS pressure in order to evaluate the contribution of the crural diaphragm to the anti-reflux function of the oesophago-gastric junction. To their surprise, they observed that, although LOS pressure was virtually abolished, gastro-oesophageal reflux was also reduced. Analysis of the underlying mechanisms showed that this was explained by a reduction in TLOSRs. The site of action of atropine is presumably central as the muscarinic antagonist, metscopolamine bromide, which does not pass the blood–brain barrier, did not reduce the rate of TLOSRs.71 Using a barostat, it was shown that atropine reduced gastric tone, resulting in an increase in intraballoon volume. As stretch is considered to be the trigger for TLOSRs, this should have increased in contrast to decreased TLOSRs, making a central action of atropine even more likely.
A similar unexpected reduction in TLOSRs was observed after the infusion of morphine. Morphine is an opioid receptor agonist known to increase LOS pressure, leading to the hypothesis that increasing nadir LOS pressure during a TLOSR could reduce reflux. Morphine had no effect on LOS relaxation during TLOSRs, but significantly reduced their frequency.70 Naloxone completely reversed the effects of morphine, suggesting that opioid μ-receptors are also involved in the triggering of TLOSRs. Although morphine certainly acts centrally, the reduction in TLOSRs by morphine might be explained by an increase in intragastric tone, thereby reducing intraballoon volume and stretch during an isobaric distension protocol.72
Recently, evidence has been provided that NO is also one of the neurotransmitters involved in the triggering of TLOSRs.68 Inhibition of NO synthesis by NG-monomethyl-L-arginine (L-NMMA) blocked the increase in TLOSRs induced by isovolumetric distension of the proximal stomach or ingestion of a meal in healthy volunteers, confirming an earlier observation in the dog.67,69 Although NO is involved in both LOS and fundic relaxation, no effects were observed on swallow-induced LOS relaxation or intragastric pressure.73,74 However, L-NMMA did affect the timing of peristalsis of the proximal oesophagus, which is primarily controlled centrally by a programmed discharge from the nucleus ambiguus.75 It seems more likely therefore that L-NMMA acts at a central or afferent level.
GABA is an important inhibitory neurotransmitter in the central nervous system, and GABA-B receptors have been identified in the vagal afferent nerve terminals and nuclei of the dorsal vagal complex (DVC).76 The GABA-B agonist baclofen, but not the GABA-A agonist muscimol, significantly reduced the rate of TLOSRs in ferrets, dogs and humans, indicating that GABA-B receptors are indeed involved in the reflex arc underlying TLOSRs.30,64,65,77 A central effect of baclofen is most likely, as a peripherally selective GABA-B agonist was ineffective. Furthermore, baclofen readily crosses the blood–brain barrier and stimulates growth hormone release. However, reductions in mechanoreceptor sensitivity and increases in gastric tone may also contribute to the effects of baclofen.78,79
An important excitatory neurotransmitter in the central nervous system is glutamate.80 Glutamate and its N-methyl-D-aspartate (NMDA) and non-NMDA (alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionate (AMPA) and kainate) receptors have been identified in the nucleus tractus solitarius and in the vagal afferents projecting to the nucleus tractus solitarius.81–83 A recent study showed that reducing glutamatergic neurotransmission using riluzole attenuated the rate of TLOSRs triggered by isovolumetric distension in healthy volunteers.66 Most likely, TLOSRs are mediated via the non-NMDA receptors, as they mediate most fast synaptic excitation during low frequency neurotransmission and are primarily involved in physiological processes.84 As drugs interacting with glutamatergic neurotransmission may have serious side-effects, such as dizziness, nausea, vomiting, diarrhoea and anorexia, their clinical use in a common disorder like GERD is probably limited.
Meal ingestion increases the rate of TLOSRs and, as meal ingestion is also associated with a rise in plasma CCK levels, CCK may also be involved in the post-prandial increase in TLOSRs. Infusion of exogenous CCK indeed increases the rate of TLOSRs in both dogs and healthy volunteers.57,67 Similarly, increased levels of endogenous CCK following a cholestyramine-enriched meal are also accompanied by increased numbers of TLOSRs.57,85 Using the CCKA receptor antagonists, devazepide and loxiglumide, reductions in TLOSRs were reported after distension of the stomach with free air, balloon, meal or meal enriched with cholestyramine, implicating an important role of CCK and the CCKA receptor in the reflex pathway underlying TLOSRs.57,59,61,85 An efferent effect is not plausible as basal LOS pressure and water swallow-induced LOS relaxation were not affected by loxiglumide. Alternatively, loxiglumide could interact with vagal afferents or within the brainstem. As intracerebroventricular administration of devazepide did not significantly modify the rate of TLOSRs, intervention with CCK neurotransmission most likely takes place at the level of the vagal afferent fibres.
The exact level at which all of these drugs interact with TLOSRs remains unknown. However, based on the pharmacological profiles of the drugs and the (lack of) effects on LOS relaxation and gastric function, it is possible to speculate on their site of action. Figure 2 depicts the possible sites of action of the various drugs. The number of neurotransmitters involved in the triggering of TLOSRs is still increasing, raising the question of whether and, if so, how these neurotransmitters interact with each other. Evidence for a potential interaction has been provided by the observation that the increase in TLOSRs during gastric distension by CCK is inhibited by NG-nitro-L-arginine methyl ester (L-NAME).67 Furthermore, various studies have provided anatomical and electrophysiological evidence for an interaction between the neurotransmitters involved in the triggering of TLOSRs, i.e. CCK, nitric oxide synthase, glutamate, GABA, opioids and acetylcholine, at the level of the enteric and/or central nervous system. These data suggest that at least some of the neurotransmitters shown to reduce TLOSRs may interact with each other. Whether these neurotransmitters are coupled in series or in parallel is not known and could be studied using combinations of different antagonists. Detailed neurochemical mapping of the nerve fibres and neurones involved in this reflex, using combined immunohistochemistry and retrograde tracer studies, will provide further insight into this complex situation.
Fundoplication is an effective surgical approach to treat gastro-oesophageal reflux symptoms and oesophagitis. During a fundoplication procedure, the gastric fundus is wrapped around the cardia for 270–360°. If a hiatal hernia is present, the stomach is repositioned in the abdomen and a crural repair performed. Ireland and co-workers showed that fundoplication virtually eliminated all TLOSRs triggered by a meal and free air (Figure 3).86,87 The mechanism by which fundoplication inhibits the triggering of TLOSRs is not known. It is hypothesized that the fundic wrap prevents stretching of the cardia region, which could be considered as the trigger zone of TLOSRs. In addition, the vagal nerve may be severed during the procedure and may explain the reduction in TLOSRs.88 It should be noted, however, that the reduction in TLOSRs and impairment of LOS relaxation might be so profound that belching is completely inhibited, resulting in bloating. Recently, a technique to perform an endoscopic gastroplasty has been developed. Using a miniature sewing device, gastric folds in the cardiac region are sewn together, resulting in a reduction in gastro-oesophageal reflux in GERD patients.89 Whether a reduction in TLOSRs contributes to the reduction in gastro-oesophageal reflux remains to be studied.
A novel anti-reflux intervention technique is endoscopic delivery of high radiofrequency energy to the submucosa of the gastric cardia. Animal studies have shown a reduction in TLOSRs after this treatment, without any effect on LOS pressure or relaxation.90 A trend towards a reduction in TLOSRs was observed in GERD patients.91 The thermal lesions created during this intervention most likely severed the vagal afferent fibres localized in the cardiac region, explaining the reduction in TLOSRs. Similarly, injection of the local anaesthetic lidocaine around the cardiac region also reduced TLOSRs, presumably by an effect on the vagal afferent fibres.92
WILL TARGETING TLOSRS REDUCE GASTRO-OESOPHAGEAL REFLUX?
From the above described studies, we can conclude that the frequency of TLOSRs can indeed be reduced by pharmacological means. Whether this will result in a reduction of gastro-oesophageal reflux is uncertain. So far, only two drugs, namely morphine and atropine, significantly decreased gastro-oesophageal reflux in GERD patients.55,70 Reductions in reflux have also been reported after loxiglumide, baclofen and L-NMMA.65,69,85 However, these results should be interpreted with caution as the studies were performed in healthy subjects, who have very low reflux rates. Finally, all of these data were derived during acute experiments, and prolonged studies evaluating the anti-reflux properties of the drugs reducing TLOSRs are still lacking. Therefore, it remains unclear whether reducing TLOSRs will indeed effectively reduce gastro-oesophageal reflux in GERD patients.
It is unlikely, however, that targeting TLOSRs will be effective in all GERD patients. In particular, patients with moderate to severe oesophagitis will probably not benefit from a reduction in TLOSRs, as other mechanisms, such as impaired oesophageal peristalsis, the presence of a hiatal hernia and absent LOS pressures, may become major determinants of gastro-oesophageal reflux. Such patients will need profound acid suppression, prokinetics and even correction of the oesophago-gastric junction with fundoplication. Therefore, patients with mild reflux disease with normal peristalsis and, ideally, with an intact oesophago-gastric junction will be the group which could potentially benefit from anti-TLOSR therapy. However, in this group also, other mechanisms can attribute to gastro-oesophageal reflux. Ideally, a ‘reflux profile’ should be obtained from a GERD patient before starting anti-TLOSR therapy. It needs to be stressed though that it remains to be determined whether this ‘reflux profile’ is consistent in time. Unfortunately, today’s ambulatory manometric systems are impractical and only used in a research setting. Furthermore, visual analysis is time consuming and validated accurate automatic analysis software is still lacking.93 Finally, non-acid factors, such as bile, may play an important role in the pathogenesis of symptoms, oesophagitis and Barrett’s oesophagus.4,94 Therefore, measurement systems capable of detecting the reflux of non-acid contents should be combined with ambulatory manometry.95,96 Although such systems have been developed, they are also primarily used in a research setting.
Most drugs inhibiting TLOSRs have considerable central and peripheral side-effects, such as dizziness and increased blood pressure.97,98 Furthermore, some of these drugs reduce TLOSRs by interacting with neurotransmission in the nucleus tractus solitarius, dorsal motor nucleus of the vagal nerve and nucleus ambiguus (NA). These nuclei not only process gastrointestinal information, but also information from the heart and lungs. Therefore, targeting TLOSRs at a central level may result in the inhibition of other vagally mediated reflexes with unexpected and even dangerous side-effects.
In summary, the inhibition of TLOSRs in humans is feasible using (ant-)agonists of the neurotransmitters involved in the reflex arc underlying TLOSRs. In addition, fundoplication and new techniques, such as the delivery of high radiofrequency energy to the submucosa of the gastric cardia, seem to intervene with the triggering of TLOSRs. Whether targeting TLOSRs is the ‘golden bullet’ in anti-reflux therapy remains to be proven, as evidence of an effective control of gastro-oesophageal reflux in GERD patients is still lacking.