Article Text
PDF
Abstract
Emerging data increasingly point towards the duodenum as a key region underlying the pathophysiology of functional dyspepsia (FD), one of the most prevalent functional GI disorders. The duodenum plays a major role in the control and coordination of gastroduodenal function. Impaired duodenal mucosal integrity and low-grade inflammation have been associated with altered neuronal signalling and systemic immune activation, and these alterations may ultimately lead to dyspeptic symptoms. Likely luminal candidates inducing the duodenal barrier defect include acid, bile, the microbiota and food antigens although no causal association with symptoms has been convincingly demonstrated. Recognition of duodenal pathology in FD will hopefully lead to the discovery of new biomarkers and therapeutic targets, allowing biologically targeted rather than symptom-based therapy. In this review, we summarise the recent advances in the diagnosis and treatment of FD with a focus on the duodenum.
- duodenal mucosa
- functional dyspepsia
- functional bowel disorder
- intestinal permeability
Statistics from Altmetric.com
Key messages
Functional dyspepsia (FD) refers to a common unexplained disorder characterised by epigastric pain and/or postprandial discomfort (fullness, early satiety).
The major FD subgroups are the postprandial distress syndrome and epigastric pain syndrome (PDS and EPS, respectively). PDS is the most prevalent subgroup but often overlaps with EPS in those who consult and frequently co-exists with gastro-oesophageal reflux and IBS, resulting in complex and under-recognised symptom patterns.
Duodenal mucosal hyperpermeability and low-grade duodenal inflammation have been associated with altered neuronal signalling and systemic immune activation in FD.
Psychiatric comorbidity such as anxiety and depression is common and may arise from gut dysfunction or immune activation, or in other cases alter gastroduodenal function potentially leading to dyspeptic symptom generation.
Recognition of subtle pathology in FD may lead to the discovery of new potential biomarkers and therapeutic targets, ultimately contributing to diagnosis and treatment.
Future classifications should include markers of underlying pathophysiology to allow targeted rather than symptom-based therapy and to identify subgroups that respond to specific interventions.
Introduction
Functional dyspepsia (FD) is a chronic GI disorder defined by upper abdominal symptoms considered to originate from the gastroduodenal region with no structural disease on routine investigation, including upper GI endoscopy.1 Two subgroups of FD were proposed by the Rome III consensus and following publication of additional supporting evidence, reiterated in the Rome IV version: postprandial distress syndrome (PDS) with postprandial fullness or early satiation, and epigastric pain syndrome (EPS) with epigastric pain or epigastric burning.2 In addition, symptoms must be severe enough to impact on usual activities with a minimal frequency of 1 (EPS) or 3 (PDS) days per week and present for 3 months with symptom onset at least 6 months before diagnosis (table 1).2 Patients with FD have a reduced quality of life and experience increased healthcare costs with loss of productivity, confirmed by a recent cross-sectional population-based study.3
Rome IV criteria for FD
A meta-analysis using a broad definition reported that the global prevalence of uninvestigated dyspepsia is 21%, with a higher prevalence in women, smokers, users of non-steroidal anti-inflammatory drugs (NSAIDs) and in those with Helicobacter pylori (Hp) infection.4 Recent figures from the USA, Canada and UK showed a 10% prevalence of FD in the adult population using the Rome IV criteria, with a similar distribution pattern across different regions (61% PDS, 18% EPS and 21% overlap).3 Whereas PDS is characterised by meal-related symptoms, epigastric pain and/or burning in EPS may be unrelated to the meal although patients often under-report meal-related pain because it may be delayed after eating.5 Although the PDS-EPS overlap group is still substantial, recognising postprandial symptoms (including epigastric pain or burning) as part of the PDS subgroup in Rome IV substantially reduced overlap compared with the Rome III definition.2 6 In FD, symptoms of upper abdominal bloating, belching, heartburn and nausea can be present, although functional nausea and vomiting, and functional bloating are categorised separately.
Symptoms of heartburn and/or IBS often co-exist with FD. Heartburn in FD may be due to pathological acid reflux or functional heartburn.7 Although functional heartburn (12%) and IBS (32%) were independent factors associated with all FD subtypes, the strongest association was found with the overlapping group.3 In the Swedish Kalixanda study, overlap of FD was present with GORD in 4%, IBS in 2% and both in 6%.8 In a follow-up study, new-onset GORD was common in FD with a higher risk in patients with duodenal eosinophilia.9 10 Moreover, a population-based study from Mayo Clinic concluded that FD is often underdiagnosed and mislabelled as GORD (box 1).11
FD subgroups and overlap
Distribution of functional dyspepsia (FD) subtypes was similar across countries (61% postprandial distress syndrome (PDS), 18% epigastric pain syndrome and 21% overlap) with a 10% prevalence in the adult population (Rome IV).
PDS is characterised by meal-related symptoms and recognising postprandial symptoms (eg, epigastric pain or burning) as part of the PDS subgroup reduces overlap.
Epigastric bloating, belching and nausea as well as weight loss are part of the dyspepsia spectrum, but vomiting suggests an alternative diagnosis.
Symptoms of heartburn and IBS are independently associated with all FD subtypes and especially the overlap group, which may be related to duodenal eosinophilia.
Translational science concepts
Brain-gut interactions
Functional GI disorders are classified by the Rome IV criteria as disorders of gut-brain interaction with contributions of both altered brain processing and luminal changes including dysbiosis (figure 1).12 13 Up to half of the subjects with uninvestigated dyspepsia in the general population identify stress as a trigger for their symptoms.14 There is evidence to support that the effect of stress may be mediated by increased permeability and immune activation.15 Recent preclinical studies indicate that stress can lead to intestinal dysbiosis, which in turn affects central nervous system function and behaviour,16 and these findings have led to the concept of the bidirectional ‘brain-gut axis’ although supporting human data are scarce. A systematic review of functional neuroimaging studies in FD concluded that several brain regions including the frontal and somatosensory cortex showed anomalies.17 Abnormal central modulation (brain to gut) and overactive visceral sensory signalling (gut to brain) may both be involved in the pathophysiology of FD, and these pathways are likely activated or modulated through psychological factors and the hypothalamic-pituitary-adrenal axis stress response.17 18 Evidence for a bidirectional interaction between central and peripheral manifestations from an Australian longitudinal study illustrated that depression without GI symptoms at baseline predicted FD, whereas anxiety and depression developed in FD patients without psychological comorbidity at baseline after 12 years of follow-up.19 Similar results were obtained in an independent population sample.20 These findings indicate an opportunity for prevention, as an average period of more than 3 years was noted before the development of GI symptoms in patients with mood or anxiety disorders.21
The gut-brain axis in FD and mechanisms of overlap with GORD and IBS. FD, functional dyspepsia; TLESR, transient lower oesophageal sphincter relaxation.
Gastric sensorimotor abnormalities
Abnormalities of gastric function including delayed emptying, impaired accommodation and hypersensitivity to distention have been reported in FD, but these changes correlate poorly or not at all with symptoms.22 Moreover, the prevalence of altered gastric sensorimotor function was similar in the Rome III-defined PDS and EPS subgroups in a large tertiary care study, suggesting a limited contribution to symptom generation.22 Although cardinal FD symptoms did not correlate with gastric emptying,23 24 a recent systematic review and meta-analysis supported associations between upper GI symptoms and gastric emptying when optimal test methods were used.25 The distinction with gastroparesis should be made as nausea may be present but vomiting is unusual in FD,2 and nausea and vomiting are significant symptoms in idiopathic and diabetic gastroparesis.26 Gastric hypersensitivity and impaired accommodation may overlap, and no association with symptoms was found for (dys-)accommodation except for severe PDS symptoms.22 27 Indeed, dyspeptic symptoms were significantly associated with psychosocial factors such as depression and somatisation and to a lesser extent with gastric sensorimotor function.28 As impaired gastric accommodation was reported in 43% of 1267 patients with functional gastroduodenal symptoms seen at a single tertiary centre and half of these patients also had abnormal gastric emptying, the study and treatment of gastric motor dysfunction likely remain relevant in at least a subset of patients.29
Emerging data increasingly point towards the duodenum as the key integrator in dyspepsia symptom generation, and it has been proposed that gastric motor dysfunction may be attributed to disordered duodeno-gastric feedback.30 31 Several studies have now reported duodenal mucosal low-grade inflammation and increased small intestinal homing T cells as markers of intestinal inflammation, which correlated with delayed gastric emptying and dyspeptic symptom severity.32 33 Evidence for the role of duodenal inflammation in neuronal signalling in FD comes from a study in Leuven showing correlations between duodenal eosinophils or mast cells and calcium transient amplitudes to high K+ or electrical stimulation.34 The exact mechanism through which duodenal inflammation induces neuronal hyperexcitability is still incompletely understood, but in analogy to IBS, a role for histamine and proteases—among others—is hypothesised. Therefore, gastric disturbances could be secondary and not primary in FD.35 Also, altered duodeno-gastric reflexes and impaired gastric accommodation may underlie co-existing reflux symptoms through increased transient lower oesophageal sphincter relaxations leading to new-onset GORD, as heartburn was found to develop more often in patients with duodenal pathology including duodenal eosinophilia.10 36 Moreover, it has also been suggested that overlapping IBS symptoms in FD could be explained by more extensive intestinal inflammation with alteration of neuronal signalling and visceral hypersensitivity37 (figure 1).
Duodenal barrier defect and immune activation
The duodenum has emerged as a key player in both GI and metabolic diseases as it regulates the passage of food as chyme from the stomach to the small intestine, where nutrients are absorbed.38 Autocrine and paracrine mechanisms in the duodenum are also involved in the mucosal defence to acid and the luminal digestion of nutrients with secretion of bile and pancreatic juice.39 Activation of duodeno-gastric feedback mechanisms by chemical or mechanical triggers influences gastric emptying, with an important role of intestinal or pancreatic hormones including incretins and gastric orexigenic hormones, which also signal to the brain.40 Besides nutrient sensing, transmucosal passage of luminal content is likely in the proximal small intestine due to regional variation in the mucosal barrier along the GI tract and the crypt-villus axis, ranging from >20 angstrom at the crypt base to 5 angstrom at the villus tip in the proximal small intestine, which is also the most permeable region with the largest intercellular pores of the GI tract.41 Although the small bowel microenvironment with a loose glycocalyx (which covers the brush border on the apical surface of the epithelial cells) allows for a closer interaction between the lumen and the host cells in comparison with the colon, the defence against potential noxious luminal factors is reflected in the barrier function and distribution of immune cells42 (figure 2). Ussing chamber experiments have shown an increased duodenal mucosal permeability in FD with a lower transepithelial electrical resistance and higher paracellular passage of fluorescent-labelled dextran molecules.32 The expression of particular cell-to-cell adhesion proteins, including tight junctions (ZO-1 and occludin), adherins junctions (β-catenin and E-cadherin) and desmosomes (desmoglein-2), was also significantly decreased and correlated with duodenal hyperpermeability and the severity of low-grade mucosal inflammation, including duodenal mast cells and eosinophils.32 This study of FD patients from Leuven (Belgium) confirmed the initial reports on duodenal eosinophilia in adults from Sweden43 and activation of duodenal eosinophils in children in an uncontrolled pilot study from the USA.44 Studies using duodenal mucosal impedance measurement also showed lower values in FD patients although the comparison with controls did not reach statistical significance after adjusting for potential confounders (eg, NSAIDs).45 Mucosal impedance values were also positively correlated with ZO-1 (tight junction protein) and inversely correlated with tissue interleukin (IL)-1b expression, implicating the role of low-grade inflammation in duodenal barrier dysfunction.45 However, challenges remain in the interpretation and reproducibility of the experimental methods. One of the limitations of the ex vivo studies is that only the integrity of the epithelial layer is measured, lacking both the superficial mucus layer and innervation by the enteric nervous system, which also affect mucosal barrier function.
The duodenal microenvironment in health and FD with pathophysiological mechanisms. In health (left), duodenal luminal content is separated from the mucosal immune system by an intact duodenal barrier. In case of duodenal barrier dysfunction such as observed in FD (right), luminal triggers from altered duodenal content and/or dysbiosis may lead to both local and systemic immune activation with disturbed signalling of duodeno-gastric feedback mechanisms by neuronal changes in afferent nerves. These changes in the duodenal microenvironment may generate dyspeptic symptoms. Arrows (red) indicate impaired duodenal mucosal integrity and eosinophil-mast cell signalling. FD, functional dyspepsia.
The local and systemic inflammatory changes in FD were recently summarised in two systematic reviews46 47 (table 2). Mast cells have been implicated in the pathophysiology of IBS and stress-induced small intestinal hyperpermeability, measured with the urinary lactulose-mannitol ratio in healthy students undergoing psychological stress during an oral examination.15 The finding of duodenal mast cells in the Leuven cohort may be explained by overlapping FD and IBS, as mast cells in the duodenum were characteristic of IBS in another study.48 The finding of duodenal eosinophilia49–51 was accompanied by cytotoxic T cell52 or mast cell32 34 50 53 infiltration in a subset of studies. Eosinophil-mast cell signalling may also be involved in stress-induced hyperpermeability as administration of corticotropin-release hormone, which can activate specific receptors on both eosinophils and mast cells,54 increased small intestinal permeability, and this effect was blocked by pretreatment with the mast cell stabiliser disodium cromoglycate (DSCG).15 Increased duodenal mucosal eosinophil counts have been associated with postprandial symptoms in Australian adults55 and children.56 The relation between duodenal eosinophilia and meal-related symptoms may be explained by altered duodeno-gastric reflexes and gastric dysaccommodation in response to the meal, although studies directly correlating duodenal abnormalities and gastric sensorimotor dysfunction are scarce.48 55 Recently, duodenal hyperpermeability (measured in Ussing chambers) was correlated with gastric emptying to solids in an FD patient cohort from Leuven (Belgium), supporting the hypothesis of altered duodeno-gastric feedback in symptom generation.57 Functional and structural submucosal neuronal changes have indeed been reported in the duodenum of FD patients, correlating with accumulation of eosinophils and mast cells in close proximity to neurons.34 Activation and degranulation of eosinophils in the presence of normal numbers may also result in neuronal changes with dyspeptic symptom generation.44 58 59 Although systemic markers of inflammation were elevated and linked with both duodenal hyperpermeability and gastric emptying as well as dyspeptic symptoms,33 57 local hypersensitivity in the duodenum could also be implicated in FD.60 61
Mucosal and systemic immune activation in FD
The most attractive hypothesis for symptom generation is loss of mucosal integrity as a primary event, which could lead to immune activation through antigen presentation with a predominant T-helper type 2 (Th2) or a Th17 response leading to recruitment and degranulation of eosinophils, triggering visceral hypersensitivity and altered motor control1 (figure 2). Eosinophils are critical effector cells of a Th2 response characterised by Th2 allergic inflammation.62 The consistent finding of duodenal eosinophilia may help explain the observed link with atopy in a UK-based primary care registry of functional GI disorders including FD63 and also autoimmune disease.64 The association with allergy may indicate a hypersensitivity reaction in some patients with early satiety.65 Also, associations of dyspeptic symptoms in the general population with herbivore but not carnivore pets and antibiotics suggest the involvement of microbiota-related components in the gut although this was not tested and further studies should control for allergy and other atopic conditions.66 Thus, a luminal trigger (food, microbiota or secreted endogenous mediators) in the duodenum may cause the observed mucosal changes in FD patients.
Duodenal acid, duodenal bile and dysbiosis
Although the causes of the barrier defect and immune activation in FD are still unknown, likely candidates include duodenal acid, bile, stress and food or microbial components. Duodenal acid perfusion resulted in delayed gastric emptying, impaired accommodation and hypersensitivity to distension in healthy subjects, suggesting a potential role for duodenal acid exposure and alterations in gastric sensorimotor function although these perfusion experiments do not reflect physiological changes occurring in the distal duodenum.67–69 Although gastric acid secretion in FD is normal, an increased duodenal acid exposure, possibly due to delayed duodenal acid clearance, has been reported with acid perfusion experiments in FD patients.60 70 Moreover, duodenal acid perfusion in healthy subjects resulted in both gastric relaxation and duodenal mucosal hyperpermeability and mast cell activation, which were however not reversed by pretreatment with DSCG.71 The role of food in triggering FD symptoms and its role in duodenal low-grade inflammatory changes have remained unexplored to date. Duodenal hypersensitivity to lipids has been reported in FD,30 with modulation of upper GI symptoms via cholecystokinin signalling in response to fat.61 The role of gluten in FD has not been investigated although symptoms may overlap with non-celiac gluten sensitivity, in which duodenal eosinophils may also cause gastric sensorimotor dysfunction (box 2).72
New observations in the duodenum
Duodenal changes including mucosal hyperpermeability and low-grade inflammation have been associated with altered neuronal signalling and systemic immune activation.
Immune activation has been observed, and patients with functional dyspepsia may have an allergic or atopic background with a Th2-predominant immune profile.
The cause of the barrier dysfunction is unknown and likely candidates include duodenal acid, bile salts, food components, stress and the mucosal microbiome.
Gastric dysfunction may be driven by impaired duodeno-gastric feedback mechanisms through local signalling or changes in systemic immune activation.
The release of bile salts in the duodenum has been implicated in the onset or worsening of dyspeptic symptoms after a meal.73 Bile salts and bacteria have a bidirectional relationship since specific bile salts have antimicrobial effects against selected bacteria and bacteria are responsible for bile salt transformation, mainly in the colon with reabsorption and reconjugation of primary and secondary bile salts before excretion in the duodenum.74 Although bacterial deconjungation may also occur in the small bowel, deconjugated bile salts were only reported at or below the detection limit in nasoduodenal aspirates from FD patients with decreased fasted concentrations of primary bile salts.75 Moreover, a decreased ratio of primary to secondary bile salts was found during fed state in FD patients compared with controls, correlating with duodenal mucosal permeability although secondary bile salt concentrations were similar between groups.76 These data suggest that the relative abundance of primary versus secondary bile salts, rather than their absolute concentrations, affects intestinal barrier function but conclusive evidence is lacking and more in-depth studies are needed to elucidate the pathophysiological mechanism.
Immune activation in the pathogenesis of FD is very evident in the postinfectious setting,77 with persisting changes in duodenal mucosal immune cells after the initial event and systemic immune activation in acute compared with unspecified-onset FD, indicating the inability of the immune system to recover from the triggering infectious insult.77 78 Similar to studies in IBS, antibodies to cytolethal distending toxin B, produced by Gram-negative bacteria causing acute gastroenteritis, were more common in FD and IBS/FD overlap compared with healthy controls in an Australian population-based study, suggesting possible under-recognition of postinfectious FD.79 The disruption in structural and/or functional microbial configuration, defined as ‘dysbiosis’, has been studied in both functional and inflammatory GI disorders.80 81 Culture-independent metagenomics and high-throughput 16S ribosomal RNA gene sequencing have revolutionised our understanding of the human gut microbiome or collective genome of microorganisms inhabiting the GI tract.82 However, the focus has almost been exclusively on faecal microbiota, which may not reflect the mucosa-associated microbiota (MAM).83 In addition, the duodenal microbiome has hardly been studied although the commensal gut microbiota of the small intestine play an essential role in nutrient acquisition, colonisation resistance to pathogens, immune development and epithelial barrier function.42 The combination of gastric acid, bile, digestive enzymes and rapid transit time in the duodenum may contribute to a hostile environment with lower density but greater diversity in the duodenal bacterial flora compared with the rectum.84 Although more challenging to study because of the lower bacterial abundance and difficult sampling, host-microbiome interactions could be more important in the small intestine with predominant Gram-positive aerobes in the duodenum compared with obligate anaerobes in the colon.80 Data on the duodenal MAM in FD are currently limited to one pilot study involving nine FD patients, with an increase in Streptococcus and total bacterial load in the duodenal MAM compared with controls, which correlated with meal-related symptom severity and quality of life.85 However, sample size was small with no information on medication, which is an important confounder, nor IBS overlap, which has also been associated with faecal and duodenal dysbiosis.80 84 86 87 Interestingly, the abundance of Veillonella in the duodenal MAM was negatively correlated with gastric emptying time,88 but replication and further studies are needed in order to unravel the potential role of the gut microbiota in FD.
Advances in clinical practice
Investigation and management
A systematic review and meta-analysis reported that endoscopic investigation for dyspeptic symptoms was normal in over 70% of referrals, thus leading to a diagnosis of FD.89 Although peptic ulcers and gastric cancer have become uncommon and rarer causes of dyspepsia, upper endoscopy with duodenal and gastric biopsies is still the preferred diagnostic method for the exclusion of coeliac disease and Hp infection, which may also cause dyspeptic symptoms.1 Following Rome IV criteria, the diagnosis of FD can only be made in the absence of structural disease including upper endoscopy,2 but this does not exclude the presence of microscopic changes such as duodenal eosinophilia. Based on a control group selected from 1001 community subjects in Sweden,43 a cut-off of 22 eosinophils per 5 high-power fields was proposed for adults,65 although results can also be expressed per mm2 as shown in adult patients,55 and a duodenal eosinophil count of >112 per mm2 was associated with a 33-fold increased risk of FD in a paediatric Australian cohort.56 Due to the limited availability of experimental investigations for testing duodenal barrier (eg, Ussing chambers32 or mucosal impedance studies45) and neuronal function (eg, calcium imaging34), new biomarkers from duodenal biopsies may include mucosal tight junction proteins32 or IL-1B expression.45 Nevertheless, none of these advanced tests can be recommended for routine clinical practice since their outcome does not affect the management. The routine use of gastric function testing is questionable as discussed below.
The complexity of the syndrome indicates a multifactorial origin, and the lack of effective therapies has been recognised by international societies such as the American College of Gastroenterology (ACG) and Canadian Association of Gastroenterology (CAG), which have recently issued guidelines.90 As discussed below, the commonly used therapies affect brain-gut or gastric pathways, which do however not always correlate with symptoms and only limited treatments target the underlying duodenal pathophysiology (figure 3). In addition, an increased awareness of side effects has limited the routine use of some drugs, including neuromodulators and prokinetics. Evidence for dysbiosis and overgrowth with oral species with proton pump inhibitors (PPI)91 92 has provided evidence for changes in the gut environment although the link with GI symptoms is still unclear. Moreover, although still early days, randomised controlled trials of probiotics and selective antibiotics have shown improvement in both symptoms and the microbiome of FD patients.93–95
Old and new treatments based on potential pathways and therapeutic targets in functional dyspepsia. Effective treatments are indicated with green boxes and arrows, with solid lines for evidence from randomised controlled trials and dashed lines for evidence from non-randomised controlled trials. Red boxes indicate potentially harmful side-effects of treatments. PPI, proton pump inhibitor.
Targeting brain-gut and gastric abnormalities
The efficacy of neuromodulators versus placebo (risk ratio (RR) 0.78; 95% CI 0.68 to 0.91; number needed to treat (NNT)=6) was limited to antipsychotics and tricyclic antidepressants (TCAs) in a recent systematic review and meta-analysis of 13 studies including 1241 patients.96 In the Functional Dyspepsia Treatment Trial, amitriptyline but not escitalopram was effective for the treatment of ulcer-like (painful) FD, with a threefold increased odds of reporting adequate relief of symptoms compared with placebo.97 In addition, FD patients with delayed gastric emptying did not respond to amitriptyline and this was not related to a treatment-induced delay in gastric emptying, indicating an effect on visceral sensitivity and abdominal pain rather than gastric motility.97 98 However, little is known on the longer-term efficacy of neuromodulators as the longest treatment duration was 12 weeks.96 The tetracyclic antidepressant mirtazapine significantly improved PDS symptoms, quality of life, nutrient tolerance and body weight in a randomised trial of 34 FD patients with significant weight loss (>10% of original body weight) and no depression or anxiety, which could be linked to the antagonism of histamine-1, adrenergic α2 and both serotonin-2C and serotonin-3 receptors.99 Finally, treatment with buspirone, a serotonin-1A receptor agonist, which relaxes the proximal stomach in healthy individuals, significantly reduced both overall and PDS-type symptoms in FD and improved gastric accommodation.100 Although the solid gastric emptying rate was unchanged, a delay in liquid gastric emptying was noted but the relevance for dyspeptic symptoms is probably limited.100
Prokinetics enhance gastric emptying, and although delayed gastric emptying and impaired gastric accommodation may be more common in PDS patients, this was not confirmed in a large study that showed a similar prevalence of gastric motor abnormalities in Rome III-defined FD subgroups.22 Moreover, a systematic analysis of 34 studies showed no evidence of a correlation between the acceleration of gastric emptying and symptom improvement in the treatment of diabetic and idiopathic gastroparesis.101 Delayed gastric emptying may be present in about 23% of FD patients, and prokinetics also affect gastric accommodation and sensitivity, which were disturbed in 37% of FD patients.22 These effects may explain the significant benefit of prokinetics in reducing ongoing dyspeptic symptoms (RR 0.81; 95% CI 0.74 to 0.89; NNT=7) in a recent meta-analysis of 29 studies involving 10 044 FD patients.102 However, significant heterogeneity and publication bias were noted and the overall effect was less convincing after removing cisapride from the meta-analysis (NNT=12), which was withdrawn from the market due to cardiac adverse events.102 Other prokinetics such as domperidone and acotiamide are not widely available, and extrapyramidal side effects (for dopamine receptor antagonists) and QTc prolongation with domperidone limit their chronic use, leading the ACG/CAG guidelines to recommend TCA in PPI-refractory FD patients before prokinetics.90
Acid-suppressive and anti-inflammatory therapies
Acid-suppressive therapy with PPIs is recommended as first-line treatment for FD by the ACG/CAG guidelines90 and the National Institute for Health and Care Excellence (NICE)103 after eradication of Hp or in Hp-negative patients. Empirical PPI therapy before a test-and-treat approach for Hp has also been proposed in cases of low (<20%) Hp prevalence.104 In a recent Cochrane meta-analysis including 6172 patients, a reduction of global dyspeptic symptoms (RR 0.88; 95% CI 0.82 to 0.94; NNT=11) with similar quality of life on PPI compared with placebo was reported.105 This was confirmed by the meta-analysis of the ACG/CAG90 (NNT=10) with no differences for different doses or types of PPI.90 105 According to the Rome IV consensus, PPIs are considered ineffective for the PDS subgroup,2 which was based on the results from an older meta-analysis including showing efficacy of PPIs in the epigastric pain but not the dysmotility-like FD subgroups according to Rome I and II criteria.106 However, a tendency for a higher efficacy of PPI in PDS (RR 0.89; 95% CI 0.77 to 1.03) versus EPS (RR 0.99; 95% CI 0.76 to 1.28) was shown in the recent Cochrane meta-analysis although the number of included ‘pure EPS’ patients was low.105 While the exact mechanism of action in PDS is unknown, PPIs are beneficial for overlapping GORD symptoms and may also reduce duodenal acid exposure.30 In addition, anti-inflammatory effects such as the downregulation of eotaxin gene transcription seen in eosinophilic oesophagitis may also reduce duodenal eosinophilia as recently shown in a cross-sectional study,107 although this requires confirmation in a prospective setting, using the Rome IV-defined subgroups.
Acid suppression with histamine-2 receptor antagonists (H2RA) is also possible with a lower NNT of 7 compared with placebo-controlled studies of PPI according to a Cochrane review from 2006 with inclusion of older studies.108 In contrast, the recent Cochrane meta-analysis also included studies directly comparing PPI and H2RA with no difference between both treatments (RR 0.88; 95% CI 0.74 to 1.04).105 However, due to the lack of high-quality comparative trials and superior antisecretory effect of PPIs, the ACG/CAG guidelines suggest PPI over H2RA as first-line therapy,90 although both are proposed in the NICE guidelines.103 Anti-inflammatory effects of H1RA due to blockade of histamine have been studied in the generation of visceral hypersensitivity in IBS109 and may also prove beneficial in FD patients with duodenal mast cell infiltration. Moreover, combined histamine-1 and histamine-2 receptor blockade with the mast cell stabiliser cromoglycate in non-responders led to an overall response rate of 90% in dyspeptic children with duodenal eosinophilia who previously had failed on H2RA.110 The beneficial effect of combined histamine-1 and histamine-2 receptor blockade was also reported in a recent retrospective case series from Australia with a trend for higher duodenal eosinophil counts at baseline in responders versus non-responders.111 Treatment with the leucotriene-1 receptor antagonist montelukast resulted in a 62% clinical response rate compared with 32% on placebo (p<0.02). In another paediatric cohort,112 the short-term clinical response of montelukast was not associated with changes in eosinophil density or activation.113
Targeting duodenal dysbiosis
Although the ACG/CAG and NICE guidelines advice against dose escalation in case of insufficient improvement with a one time per day dose of PPI during 4–8 weeks, inappropriate and prolonged use of PPIs even in the absence of clinical benefit is frequently reported. Moreover, PPIs have been associated with an increased risk of enteric infections (including Clostridium difficile)114 and overabundance of oral or potentially pathogenic flora in the gastric115 and faecal microbiome.91 Data on the duodenal microbiota are still lacking. The increase in oral or potentially pathogenic flora on PPI was represented by Streptococcus in the gastric MAM, and this may have contributed to persisting dyspeptic symptoms in a subgroup of patients.116 Therefore, this study and others have questioned the use of PPI therapy in FD patients, as does their limited efficacy.90 These limited data suggest that stopping PPIs in patients without a clear benefit may be a first option to target dysbiosis in FD. Eradication therapy with antibiotics improves symptoms in Helicobacter pylori-related dyspepsia (HpD), and the effect was greater with metronidazole and in patients with microscopic duodenitis,117 118 suggesting an effect on the duodenal microbiome, which is distinct from Hp eradication per se.119
Probiotics may improve HpD by inhibition of Hp, but their efficacy was also demonstrated in Hp-uninfected FD patients.93 94 Following a 12-week period of daily yoghurt ingestion containing 109 colony-forming units of Lactobacillus gasseri OLL2716 (LG21) or placebo (fermented milk product without L. gasseri) in Hp-negative FD patients, a trend for a positive overall effect on gastric symptoms (p=0.07) and significantly higher elimination rates for the major FD symptoms including PDS-like but not EPS-like symptoms was noted.93 In a follow-up study, changes in the gastric fluid microbiota were reported in FD patients at baseline with a restoration after intake of LG21.94 In addition, bile acids were more frequently detected in gastric fluid samples from FD patients compared with controls and although bile acids did not decrease with intake of LG21, the presence of an intestinal-like bacterial profile was restored after probiotics.94 This study and others have suggested the presence of small intestinal bacterial overgrowth (SIBO) in a subset of FD patients,94 120 which could be a marker and target although no standardised diagnosis is available.
Treatment with rifaximin, a non-absorbable and selective antibiotic with established efficacy for the treatment of abdominal pain and bloating in non-constipated IBS,121 may reverse SIBO as measured with hydrogen breath testing,122 but this putative mechanism is highly controversial. In a recent randomised trial, rifaximin was superior to placebo for the adequate relief of both global dyspeptic symptoms (78% vs 52%) at 8 weeks and postprandial fullness, bloating and belching at 4 weeks in FD patients who were Hp negative.95 The effect on symptoms was more pronounced in women with a similar incidence of adverse effects in the rifaximin and control group. Although the exact mechanism of action is not known, the antibiotic effect and increased solubility of rifaximin with bile salts could explain its therapeutic effect in FD patients with changes in the duodenal microenvironment (box 3).123
Targeted treatments in FD
Treatments for brain-gut and gastric dysfunction, including neuromodulators and prokinetics, have a limited efficacy and are hampered by side effects.
Proton pump inhibitors may have both acid-suppressive and anti-inflammatory effects in functional dyspepsia (FD) patients with duodenal eosinophilia and impaired barrier function.
Targeting duodenal inflammation and the eosinophil-mast cell axis with mast cell stabilisers and antihistamines requires studies in adult populations.
Duodenal dysbiosis in FD may be treated with selective antibiotics, such as rifaximin, but requires confirmation in Western populations.
Conclusion
The paradigm of FD as the most common functional upper GI disorder with ‘no evidence of structural disease’ has recently been challenged.32 33 43 56 78 Despite the common occurrence of FD in up to 15% of the general population, the underlying pathophysiology remains unclear.1 2 30 Emerging data point towards the duodenum as the key integrator in symptom generation. Meal-related symptoms in the PDS and overlap group are linked with increased duodenal mucosal permeability and low-grade inflammation with submucosal neuronal changes,32 34 43 55 providing a basis for postprandial dyspeptic symptoms and a potential diagnostic or predictive tool. However, it is still unclear whether these changes are indeed causal or secondary to an unknown cause. The association with GORD due to gastric disturbances secondary to duodenal inflammation and the possibility of overlapping IBS depending on the extent of intestinal inflammation also require further study.10 Current treatment options are of limited efficacy and target symptoms and gastric sensorimotor function rather than the underlying duodenal pathology.35 37
Future classifications of FD subtypes should take into account the different pathophysiological mechanisms in order to allow targeted and not just symptom-based therapy. Recognition of immune activation will hopefully lead to the discovery of new biomarkers and therapeutic targets, ultimately contributing to the diagnosis and treatment of this chronic and challenging GI condition. The effect of acid suppression on the microbiome is highly relevant as PPIs are frequently scrutinised because of overprescribing and potential long-term adverse events. PPI-induced dysbiosis may also help explain the limited efficacy of long-term PPIs as the current first-line treatment in FD90 and the possibility of persisting dyspeptic symptoms in a subgroup of patients on PPIs.117 Other treatments targeting the microbiome in FD look promising but require confirmation.
References
Footnotes
Contributors LW drafted the manuscript. JT revised the manuscript. MMW revised the manuscript. NJT drafted and revised the manuscript. TV drafted and revised the manuscript.
Funding LW and TV are supported by the Flanders Research Foundation (FWO Vlaanderen) through a doctoral fellowship and a senior clinical research mandate, respectively. JT is supported by a Methusalem grant of KU Leuven. NJT is supported by an NHMRC Level 3 Investigator Grant and is lead investigator of an NHMRC Centre of Research Excellence.
Competing interests NT: Consultancy: Allergan, IM Health Sciences, Takeda, Theravance, Danone, Sanofi. JT: Consultancy: AlfaWassermann, Allergan, Danone, Shire, Takeda, Theravance, Tsumura and Zeria Pharmaceuticals; Speaker fees: Abbott, Allergan, Shire, Takeda and Zeria Pharmaceuticals. TV: Speaker fees: Abbott. Research Grant: Danone.
Patient consent for publication Not required.
Provenance and peer review Commissioned; externally peer reviewed.