Sawada A, Guzman M, Nikaki K, Sonmez S, Yazaki E, Aziz Q, Woodland P, Rogers B, Gyawali CP, Sifrim D. Identification of Different Phenotypes of Esophageal Reflux Hypersensitivity and Implications for Treatment.Clinical Gastroenterology and Hepatology,2020
Identification of Different Phenotypes of Esophageal Reflux Hypersensitivity and Implications for Treatment
Akinari Sawada, MD, PhD1, Mauricio Guzman, MD1, Kornilia Nikaki, MD1, Shirley Sonmez, MRS1, Etsuro Yazaki, PhD1,Qasim Aziz, MD, PhD1, Philip Woodland, MD, PhD1, Benjamin Rogers, MD2, C. Prakash Gyawali, MD, PhD2, Daniel Sifrim, MD, PhD1
Background & Aims: Reflux hypersensitivity (RH), a functional esophageal disorder, is detected in 14%–20% of patients who present with typical esophageal symptoms. As many as 40% of patients with RH do not respond to treatment with pain modulators or proton pump inhibitors (PPIs); behavior disorders might contribute to lack of treatment efficacy. We aimed to assess the prevalence of behavioral disorders and their effects on typical reflux symptoms in patients with RH.
Methods: We performed a retrospective study of 542 patients with PPI-refractory esophageal symptoms (heartburn, regurgitation, or chest pain) or with symptoms that responded to PPI therapy, evaluated for anti-reflux surgery from January 2016 through August 2019 at a single center in London, United Kingdom. We collected data on symptoms, motility, and impedance-pH monitoring and assigned patients to categories of RH (n=116), functional heartburn (n=126), or non-erosive reflux disease (n=300).
Results: Of the 116 patients with a diagnosis of RH, 59 had only hypersensitivity, whereas 57 patients (49.2%) had either excessive supragastric belching (SGB, 39.7%), based on 24-hour impedance-pH monitoring, or rumination (9.5%), based on postprandial manometry combined with impedance. The prevalence of SGB and rumination in patients with RH was significantly higher than in patients with functional heartburn (22%; P<.001). Patients with RH and rumination were significantly younger (P=.005) and had the largest number of non-acid reflux episodes (P=.023). In patients with RH with SGB, SGB episodes were associated with 40.6% of marked reflux symptoms (heartburn, regurgitation, or chest pain), based on impedance-pH monitoring. In patients with RH and rumination, 40% of reflux-related symptoms (mostly regurgitation) were due to possible rumination episodes.
Conclusions: Almost half of patients with a diagnosis of RH have behavior disorders, including excessive SGB or rumination. Episodes of SGB or rumination are associated with typical reflux symptoms. Segregation of patients with diagnosis of RH into those with vs without behavioral disorders might have important therapeutic implications.
KEY WORDS: NERD, FH, psychologic, pain perception
Need to Know
Background: Many patients with reflux hypersensitivity do not respond to treatment with pain modulators or proton pump inhibitors — behavior disorders might contribute to lack of treatment efficacy.
Findings: Almost half of patients with a diagnosis of reflux hypersensitivity have behavior disorders, including excessive supragastric belching or rumination, which are associated with typical reflux symptoms.
Implications for patient care: Patients with reflux hypersensitivity and behavior disorders might require different therapeutic strategies than patients without behavior disorders.
Gastroesophageal reflux disease (GERD) is defined as reflux of gastric contents causing troublesome symptoms and/or complications.1 Based on symptom profile, endoscopic findings and distinct patterns on ambulatory reflux monitoring, patients with reflux symptoms can be phenotyped into Erosive Reflux Disease, Non-Erosive Reflux Disease (NERD), Reflux Hypersensitivity (RH) and Functional Heartburn (FH).
Hypersensitivity to acid reflux episodes on ambulatory pH monitoring was described many years ago,2 which has expanded to identification of hypersensitivity to non-acid reflux episodes with impedance-pH monitoring.3 According to Rome IV criteria, the diagnosis of RH requires (i) occurrence of heartburn or chest pain, (ii) normal endoscopy, (iii) absence of major esophageal motility disorders and (iv) normal acid esophageal exposure but positive reflux symptom association (RSA).4 In patients evaluated for heartburn, RH has a prevalence of around 14%.5 Taking into account the ~28% prevalence of GERD,6 the prevalence of RH is significant, especially in PPI-refractory patients.7-9
Since RH symptoms are time-related to reflux episodes, treatment recommendations include increasing acid suppression7,10 and modulation of pain perception4, but approximately 40 % remain refractory to these approaches.9,11,12 While the reason for refractoriness is not completely understood, psychological and behavioral disturbances are increasingly recognized within esophageal disorders.13 For example, a significant proportion of PPI-refractory patients have postprandial rumination or increased supragastric belching (SGB)14-16, which do not respond to PPIs or pain modulators.17,18 We hypothesized that undiagnosed behavioral disorders might contribute to treatment failure in RH patients. The aim of this study was to re-assess symptom profiles and impedance pH tracings in a large cohort of RH patients to evaluate for behavioral disorders.
We identified patients with PPI-refractory esophageal symptoms (heartburn, regurgitation or chest pain) or PPI-responsive patients evaluated for anti-reflux surgery, by interrogating the electronic database (January 2016-August 2019) at the Royal London Hospital GI Physiology Unit. Patients were included if they were over 16 years old, underwent high resolution manometry (HRM) and off-PPI impedance-pH monitoring. Patients were excluded if they had (i) endoscopic esophagitis, Barrett’s esophagus or eosinophilic esophagitis, (ii) HRM diagnosis of major esophageal motility disorders, or (iii) belching as the main symptom. Medication use and prior foregut surgery were not exclusionary. Consequently, 10 patients with persistent symptoms on antidepressants (Citalopram and Gabapentin (N=1), Gabapentin (N=2), Citalopram (N=2), Amitriptyline (N=2), Nortriptyline (N=1), Venlafaxine (N=1), and Sertraline (N=1)), opioids (N=9), prior anti-reflux surgery (N=4) and prior bariatric surgery (N=1, sleeve gastrectomy) were included in the RH group.
Using Lyon consensus acid exposure time (AET) thresholds,19 we divided patients into those with normal AET (< 4%) and pathological AET (> 6%). For clarity of interpretation, patients with borderline AET (4-6%) were excluded from the analysis. To reliably evaluate RSA, patients were included only if they had ≥3 GERD symptoms during impedance-pH monitoring; RSA required both symptom index (SI) and symptom association probability (SAP) to be positive.19,20
For this retrospective analysis of clinically indicated tests with no identifiable patient data, formal ethics approval was not deemed necessary, but we obtained approval from our Quality and Service Improvement department at the Royal London Hospital.
Questionnaire, high resolution manometry and ambulatory impedance-pH monitoring
As per our clinical routine, all patients completed the Reflux Disease Questionnaire (RDQ)21 prior to HRM testing.
HRM (Sandhill Scientific, Highlands Ranch, CO, or ManoScan, Medtronic, Minneapolis, MN) and impedance-pH monitoring (Sandhill Scientific, Highlands Ranch, CO, USA, or OMOM System, Jinshan Science and Technology, Chongqing, China) were performed after overnight fasting. An HRM catheter with 36 solid-state pressure sensors spaced 1 cm apart was inserted transnasally to record pressures from the stomach to the upper esophageal sphincter. After catheter placement and identification of the lower esophageal sphincter (LES), a 30-second baseline recording and ten 5ml water swallows were performed in the right lateral position to assess esophageal motility. When rumination was suspected on medical interview, HRM combined with impedance (HRM/Z) (Sandhill Scientific, Highlands Ranch, CO) was performed with postprandial evaluation for at least 15 minutes.
Ambulatory impedance-pH monitoring was performed off-PPI following HRM or HRM/Z, with PPIs and/or H2 receptor antagonists discontinued for at least 7 days. The impedance-pH catheter was inserted with the esophageal pH sensor positioned 5cm above the LES, and 6 impedance channels positioned 3, 5, 7, 9, 15 and 17cm above the LES respectively, and connected to a portable recorder.
HRM and impedance-pH monitoring studies were re-analyzed and manually edited using ManoView 3.0 (Medtronic), Bioview Analysis (Sandhill Scientific) and/or OMOM Analysis (Jinshan Science and Technology) specifically for the purpose of this study.
High resolution manometry
Major motility disorders on HRM included achalasia, esophagogastric junction (EGJ) outflow obstruction, distal esophageal spasm, hypercontractile esophagus and absent contractility, assessed using Chicago classification v3.0.22 EGJ morphology and EGJ-contractile integral (EGJ-CI) were analyzed as recently described.23
Ambulatory impedance-pH monitoring
Automated analysis of impedance-pH tracings was followed by manual editing of reflux episodes using published criteria24 and measurement of Mean Nocturnal Baseline Impedance (MNBI).25
Symptoms occurring within 2 minutes following onset of reflux were considered associated with the reflux episode (symptomatic reflux). Symptom index (SI) and symptom association probability (SAP) were used to assess RSA.26,27 SI measures the proportion of symptoms associated with reflux episodes during the 24 hr recording, positive when ≥50%. SAP evaluates if reflux episodes and symptoms co-occurred purely by chance, by assessing for the presence or absence of reflux and/or symptoms for each 2 min segment of the ambulatory reflux study. Using a Fisher exact test on this data, p<0.05 indicates the probability of chance association is <5%, corresponding to a positive SAP (>95%).
Definitions of Non-erosive reflux disease, Reflux Hypersensitivity and Functional Heartburn
NERD, RH and FH were defined using Rome IV criteria.4 Although originally intended only for heartburn and chest pain, we adapted RH criteria to include regurgitation, since the Montreal definition describes regurgitation as a typical GERD symptom,1 and several RH studies have similarly included regurgitation.9,11,12 All patients had normal endoscopy. NERD was diagnosed when AET was >6% regardless of RSA, RH required AET<4% with positive RSA, while FH required AET<4% with negative RSA.
Definitions of supragastric belching (SGB) and rumination
SGBs were identified on impedance-pH monitoring using previously described criteria,28 consisting of an abrupt ≥1000 Ω antegrade rise in impedance, followed by retrograde recovery to baseline. Liquid reflux episodes occurring within 1 s following SGB defined SGB-induced reflux. Based on our previous analysis of prevalence of SGB in healthy asymptomatic subjects, >13 SGB episodes/24 hours were considered pathological.29
Rumination was diagnosed using HRM/Z as previously described.30 Based on HRM/Z findings, a patient was considered a ruminator even in the presence of SGB on impedance-pH monitoring. A rumination episode was defined as impedance-detected retrograde gastroesophageal liquid flow reaching the proximal esophagus, associated with a rapid gastric pressure increase (>30 mmHg). Rumination was characterized using HRM/Z as follows: (i) primary rumination: rumination following abdominal pressure increase; (ii) secondary rumination: rumination occurring during spontaneous gastroesophageal reflux; (iii) SGB-associated rumination: rumination following SGB.30
Relationship between SGB or possible rumination episodes and GERD symptoms during impedance-pH monitoring
In patients with RH, we analyzed the association between SGB or possible rumination episodes and GERD symptoms (i.e. heartburn, regurgitation or chest pain). We considered a GERD symptom to be triggered by SGB if marked by the patient within 20 seconds following SGB. We selected a short time window because we expected very rapid perception of SGB-induced esophageal distention. Three patterns were identified: (i) symptom associated with SGB without liquid reflux (Figure 1A); (ii) symptom associated with SGB-induced reflux (Figure 1B); (iii) symptom associated with SGB occurring during reflux (Figure 1C).
Figure 1. Association between SGB and GERD symptoms. During impedance-pH monitoring, GERD symptoms were marked within 20 seconds after either SGB without liquid reflux (A), SGB causing reflux (B) or SGB during liquid reflux (C). SGB was associated with 40.5% (24.5–66.7) of GERD symptoms (D). SGB; supragastric belching, GERD; gastroesophageal reflux.
* p=0.022 compared to SGB which caused refluxes
Non-acid reflux episodes with high proximal extent (reaching the most proximal impedance channel, 17 cm above LES) within the first postprandial hour were regarded as possible rumination episodes.15 We calculated the proportion of early postprandial GERD symptoms that could be considered possible rumination episodes. In type 3 rumination, SGB-induced reflux with high proximal extent occurring outside 1-hour postprandial periods were also included in the calculation.
Continuous variables are expressed as median (interquartile), while categorical variables are expressed as numbers (percent). Overall differences across three groups were assessed using theKruskal-Wallis test for continuous variables, or Fisher exact test for categorical variables. Mann-Whitney test and Fisher’s exact test were used to compare between each pair of the three groups. To allow for multiple testing within each variable, the p-values from the pairwise comparisons were given a Bonferroni adjustment. Univariate logistic regression was performed to identify predictors of SGB or rumination in patients with RH, using odds ratios (OR) and 95% confidence interval (CI) for predictors identified on univariate analysis, followed by multivariate analysis using predictors with P<0.20 on univariate analysis. All analyses were performed using R software, version 3.3.1 (R Core Team, Vienna, Austria). P value < 0.05 was considered statistically significant.
Of 597 patients, 55 were excluded (38 had a major motility disorder, 10 had technical problems with impedance-pH tracings, and 7 had very low MNBI (<1000 ohms) from significantly impaired mucosal integrity). The final study cohort included 116 patients with RH, 126 patients with FH, and 300 patients with NERD. Demographics and clinical characteristics are described in Supplementary Table 1.
Prevalence of SGB and rumination
The proportion of patients with excessive SGB in the RH group (39.7%) was significantly higher than in the FH group (22%, P=0.01) and similar to the NERD group (37.7%, P=1). On the other hand, rumination was detected on HRM/Z in 11 patients (9.5%) with RH, which was significantly higher than in FH (0 (0%), P<0.001) or NERD groups (4 (1.3%), P<0.001). Taken together, the proportion of patients with excessive SGB or rumination was 49.2% in RH, 22% in FH and 39% in NERD (Figure 2).
Figure 2. Proportion of patients with excessive supragastric belching during impedance-pH monitoring and with rumination during postprandial HRM/Z in RH, FH and NERD. RH; Reflux hypersensitivity, FH; Functional Heartburn, NERD; Non-erosive reflux disease.
Clinical characteristics of reflux hypersensitivity (RH) patients
Within the 116 patients with initial diagnosis of RH, we identified 3 phenotypes; (i) 59 patients (51%) with “pure” RH (RH-pure), (ii) 46 patients (40%) with excessive SGB (RH-SGB), and (iii) 11 patients (9%) with rumination (RH-RUM) confirmed by HRM/Z (Table 1). Patients with RH-RUM were significantly younger than the other phenotypes (P=0.005). Proportion of heartburn and regurgitation were similar in patients with RH-pure and RH-SGB, whereas patients with RH-RUM had predominant regurgitation (Table 1).
Table 1. Characteristics of study patients tested off PPI segregated by acid exposure time (AET)
b P<0.05 compared to Supragastric belching
Percentages in brackets relate to each column
RH; Reflux Hypersensitivity, BMI; Body mass index, RDQ; Reflux Disease Questionnaire, SI; symptom index, SAP; symptom association probability
Esophageal physiology patterns within RH phenotypes (Table 2)
Esophageal body motility was similar among RH phenotypes; IEM proportions were also similar. The RH-SGB group had type II EGJ morphology more often than RH-pure and RHRUM groups. In contrast, EGJ pressures (i.e. EGJ-CI) were similar across phenotypes. Patients with RH-RUM had significantly more non-acid reflux than RH-pure and RH-SGB. MNBI measurements were normal (> 2000 ohms) in the distal channels 3 and 5 cm above the LES in all phenotypes. The RH-RUM group had significantly higher MNBI compared to the other 2 phenotypes.
Table 2. High resolution manometry and impedance-pH monitoring results in the three RH groups
b P<0.05 compared to Supragastric belching
EGJ; esophagogastric junction, EGJ-CI; EGJ-contractile integral, AET; acid exposure time, MNBI; mean nocturnal baseline impedance, LES; lower esophageal sphincter, SGB; supragastric belching.
Relationship between SGB and GERD symptoms in patients with RH-SGB during impedance-pH monitoring (Table 3)
Patients with RH-SGB had reflux symptoms (i.e. heartburn, chest pain or regurgitation). rather than belching as their main symptom. Overall, they reported 312 episodes of heartburn, 462 episodes of regurgitation and 26 episodes of chest pain.
SGBs were associated with 40.6% (25.0–66.7) of GERD symptoms (Figure 1D). SGB causing reflux (18.9% (4.4–33.3)) associated with symptoms significantly more often compared to SGB without reflux (7.0% (0.0–20.0), P=0.022) and SGB during reflux (8.0% (0.0–22.0), P=0.058) (P=0.014) (Figure 1D).
Table 3. Characteristics of study patients tested off PPI segregated by acid exposure time (AET)
Relationship between possible rumination events and GERD symptoms in patients with RH-RUM during impedance-pH monitoring
Patients with RH-RUM had a median of 9 (6–17) possible postprandial rumination episodes on impedance-pH monitoring. In these patients, 40.0% (7.8–54.5) of the symptomatic (mainly regurgitation) liquid retrograde events were considered as possible rumination episodes.
Predictive factors for presence of SGB or Rumination in patients with RH (Supplementary Table 2 and 3)
Multivariate analysis showed that abnormal EGJ morphology (type II or III EGJ) was an independent predictive factor for SGB (RH-SGB) (OR, 5.56; 95% CI, 1.64–18.9; P=0.006). On the other hand, younger age (OR, 0.83; 95% CI, 0.71–0.98; P=0.025) and larger number of nonacid reflux episodes (OR, 1.08; 95% CI, 1.02–1.15; P=0.007) were associated with RH-RUM.
Supplementary Table 2. Univariate and Multivariate analysis of predictive factors of RH-SGB in RH group
Supplementary Table 3. Univariate and Multivariate analysis of predictive factors of RH-RUM in RH group
Based on recent emphasis on behavioral disorders in PPI-refractory states14, we hypothesized that undiagnosed behavioral disorders might account for some of the 40% reported refractoriness to RH management7-9. We found that 21% of patients with PPI-refractory reflux symptoms investigated with endoscopy/reflux monitoring are initially diagnosed as having RH, and 49% of these patients have pathological SGB or rumination. Further, in RH-SGB patients, SGB triggered 34% of symptomatic refluxes and explained 41% of GERD symptoms. In RH-RUM patients, 40% of reflux symptoms were associated with possible rumination events.
As many as 24%-34% of symptomatic PPI refractory patients are reported to have RH.7-9,11 Our study demonstrated a slightly lower RH prevalence (21%), probably from our stringent definitions requiring strict AET thresholds and both SI and SAP to be positive for RSA, to reduce the likelihood of erroneous categorizing of phenotype19,20,27.
While belching is a common GERD symptom,31 pathological numbers of SGBs are identified when excessive belching is a dominant symptom. Alternatively, approximately 40% of patients with pathological SGB report reflux symptoms rather than belching16,29; Hemmink et al. also reported that 38% (9/24) of their SGB patients did not have belching as their main symptom.32 Our prevalence rate of pathological SGB (34% (187/543), (Supplementary table 1) is concordant with other studies showing a similarly high SGB prevalence in GERD patients.32
We found a high prevalence of pathological SGB in both RH and true NERD, and a proportion of SGB may contribute to increased AET in NERD.16 Alternatively, some patients diagnosed as RH during impedance-pH monitoring could in fact have NERD misdiagnosed because of day-today AET variability, and may be identified as NERD on prolonged wireless reflux monitoring.33 Of 116 patients initially diagnosed as RH, 9.5% had rumination on HRM/Z, a higher proportion than that observed in the NERD and FH groups. A higher prevalence (20%) was reported by Yadlapati et al. on postprandial HRM/impedance in PPI refractory patients,14 likely related to different study populations. We identified 3 different RH phenotypes, RH-pure, RH-SGB and RH-RUM, without many physiological differences between phenotypes. Patients with RH-SGB had hiatal hernias more often, while patients with RH-RUM had more non-acid reflux and higher MNBI, albeit within the normal range.
Central/peripheral sensitization, psychological co-morbidities and stress can contribute to RH pathophysiology.34-36 Hidden excessive SGB caused one-third of symptomatic refluxes and underlined 40% of marked reflux symptoms during impedance-pH monitoring in patients with initial RH diagnosis. This finding suggests that in RH-SGB, true hypersensitivity might be accountable for only part of the mechanisms underlying symptoms, and explains the poor response to pain modulators. Antidepressants recommended for RH management are not particularly effective for SGB or rumination,17,18 and patients with RH-SGB typically require a combination of pain modulators and cognitive behavioral therapy (CBT).4,16
Our findings have both diagnostic and therapeutic implications for patients with PPI-refractory GERD symptoms. While PPI-refractory patients should be carefully questioned for belching and regurgitation to diagnose hidden behavioral disorders, clinical history alone might not be enough to establish diagnosis of pathological SGB or rumination. Young patients with predominant postprandial regurgitation will benefit from HRM/Z using a test meal to diagnose rumination or SGB.37 While expert interpretation of impedance-pH monitoring is often required, certain tips can lead less-experienced readers to identify SGB, rumination, and the different RH phenotypes. In RH patients, more than 60% of SGB are related to reflux events; therefore, initial automatic detection followed by careful manual review of all reflux episodes can identify pathological SGB. Frequent symptomatic non-acid reflux in the early postprandial period is a characteristic of rumination episodes15. Based on our current findings, EGJ types II and III, young age and significant numbers of non-acid reflux episodes should prompt a careful analysis of the impedance-pH tracings for evidence of RH-SGB or RH-RUM. There are potential treatment implications for each RH phenotype. In pure RH, acid suppression and pain modulators are optimized,7,10 but escalation of PPI dosing or anti-reflux surgery should only be considered if a very clear association is demonstrated between acid reflux and symptoms. In RH with SGB; a dual approach using pain modulators and CBT could theoretically suffice. In RH with rumination, behavioral intervention with diaphragmatic breathing is considered the best current treatment.18
This study has some limitations. First, this is a retrospective single tertiary center study. However, the retrospective analysis helped us to identify phenotypes within a large RH cohort. It is unlikely that a prospectively collected RH cohort would provide different information, as phenotypes were identified on objective impedance pH tracing analysis. Second, it is possible that some ruminators potentially remain undiagnosed, since HRM/Z was employed based on the treating clinicians’ expectation. Finally, lack of an intervention protocol limits assessment of the clinical utility of phenotyping RH patients. In the second step of our project, a prospective outcome study will define the value of identification of the RH phenotypes in predicting treatment outcomes using CBT or pain modulators.
In conclusion, high proportions of RH patients with PPI-refractory esophageal symptoms have hidden SGB or rumination, accounting for up to 41% of symptomatic reflux. These disorders might partly explain the 40% refractoriness to PPI and pain modulators in these patients. Diagnosing SGB or rumination in patients with RH can therefore modify the therapeutic strategy. Prospective outcome studies combining psychological therapies and pain modulators are needed to prove this hypothesis.
Address all correspondence to: Daniel Sifrim MD, PhD, Upper GI Physiology Unit, Barts and The London School of Medicine and Dentistry Queen Mary, University of London, UK. Royal London Hospital, Wingate Institute of Neurogastroenterology. 26 Ashfield Street London E12AJ, United Kingdom. Phone: + 44 (0) 20 7882 2631. Fax: + 44 (0) 20 7375 2103. Email: d.sifrim@ qmul.ac.uk
Grant support: None declared.
Disclosures: Daniel Sifrim receives research grants from Reckitt Benckiser UK, Jinshan Technology China and Alfa Sigma, Italy. C. Prakash Gyawali consults for Medtronic, Diversatek, Ironwood, IsoThrive and Quintiles. The remaining authors declare no conflicts of interest with this study.
Guarantor of the article: DS. Author roles: DS and AS considered study concept and design of study protocol. AS, KN, SS, EY and DS performed and analyzed HRM and impedance-pH studies. MG, BR and CPG analyzed special metrics of HRM. AS and DS interpreted results of studies, prepared figures and drafted the manuscript. PW, QA and CPG edited and revised the manuscript with final approval from all the authors.
Назад в раздел
Популярно о болезнях ЖКТ читайте в разделе "Пациентам"
Информация на сайте www.gastroscan.ru предназначена для образовательных и научных целей. Условия использования.