Gyawali CP, Tutuian R, Zerbib F, et al. Value of pH Impedance Monitoring While on Twice-Daily Proton Pump Inhibitor Therapy to Identify Need for Escalation of Reflux Management. Gastroenterology, V. 161, Issue 5, 2021. P.1412-22.
Value of pH Impedance Monitoring While on Twice-Daily Proton Pump Inhibitor Therapy to Identify Need for Escalation of Reflux Management
C. Prakash Gyawali,1 Radu Tutuian,2 Frank Zerbib,3 Benjamin D. Rogers,1,4 Marzio Frazzoni,5
1 Division of Gastroenterology, Washington University School of Medicine, St Louis, Missouri;
2 Division of Gastroenterology, University Clinics for Visceral Surgery and Medicine, Bern University Hospital, Bern, Switzerland;
3 Centre Hospitalier Universitaire de Bordeaux, Centre Medico-chirurgical Magellan, Hôpital Haut-Lévêque, Gastroenterology Department, Université de Bordeaux; Institut National de la Santé et de la Recherche Médicale CIC 1401, Bordeaux, France;
4 Division of Gastroenterology, Hepatology, and Nutrition, University of Louisville School of Medicine, Louisville, Kentucky;
5 Department of Specialized Medicine, Digestive Pathophysiology Unit, Baggiovara Hospital, Modena, Italy;
6 Digestive Physiology, Hospices Civils de Lyon, Hôpital Edouard Herriot, Université de Lyon, Lyon, France;
7 Laboratory of Therapeutic Applications of Ultrasound, Institut National de la Santé et de la Recherche Médicale U1032, Université de Lyon, Lyon, France;
8 Division of Gastroenterology, Department of Surgery, Oncology and Gastroenterology, University of Padua, Padua, Italy;
9 Gastroenterology Unit; Department of Translational Research and New Technology in Medicine and Surgery. University of Pisa, Pisa, Italy;
10 Division of Gastroenterology, Mayo Clinic, Scottsdale, Arizona; and
11 Barts and The London School of Medicine and Dentistry Queen Mary, University of London, London, United Kingdom
Keywords: pH Impedance Monitoring; Gastroesophageal Reflux Disease; Acid Exposure Time; Reflux Episodes; Proton Pump Inhibitor.
Ambulatory pH impedance monitoring is commonly used in the investigation of esophageal symptoms in the context of gastroesophageal reflux disease (GERD)1,2 and is often performed when symptoms persist or do not improve with empiric GERD management.2 In individuals where GERD diagnosis is not firmly established, ambulatory reflux monitoring is performed while off antisecretory therapy.2 In patients with proven GERD on prior endoscopy (high-grade erosive esophagitis, Barrett’s esophagus, or peptic stricture) or on ambulatory reflux monitoring (elevated acid exposure time [AET]), further evaluation is often necessary when symptoms persist despite seemingly adequate acid suppression.1–3 Results from such testing determines if reflux management needs to be escalated (more potent acid suppression, antireflux surgery, or endoscopic antireflux interventions) or if additional symptomatic management should be undertaken (neuromodulators for functional symptoms, diaphragmatic breathing, cognitive-behavioral therapy) based on presence or absence of pathologic GERD despite therapy and/or reflux-symptom association (RSA).3
The optimal reflux monitoring study suggested for patients with proven GERD is pH impedance testing on maximal proton pump inhibitor (PPI) treatment,2,3 to explain the basis for persisting symptoms despite maximal acid suppressive therapy in patients with prior conclusive evidence of GERD.4 However, metrics and thresholds that support ongoing pathologic GERD when reflux monitoring is performed on therapy remain in question. The Lyon Consensus proposed using similar thresholds for esophageal reflux parameters for off- and on-PPI testing.2 However, the median distal esophageal AET in 135 patients with typical symptoms studied with pH monitoring on PPI was 1.0 (interquartile range [IQR], 0.2–3.8), although these patients were not confirmed to have pathologic GERD before testing.5 Moreover, there is limited evidence indicating that <40 reflux episodes on therapy is associated with improvement in GERD symptoms and satisfaction with therapy on validated questionnaires, and that >80 reflux episodes despite PPI therapy is associated with likelihood of symptom improvement with the use of invasive GERD management.6 Other metrics extracted from pH impedance monitoring (baseline impedance, post-reflux swallow-induced peristaltic wave [PSPW]) have very limited evidence to support their use in individuals studied while on PPI.7
Thus, the Lyon Consensus thresholds may be too high to have optimal sensitivity for recognition of refractory acid burden while on PPI, whereas previous published thresholds using pH-metry alone potentially suffer from lack of specificity.5 We evaluated pH impedance studies performed while on PPI therapy in healthy volunteers and in patients with proven GERD, to determine optimal metrics and thresholds to better define whether or not persistent symptoms on PPI are due to pathologic reflux, and to determine need for escalation of antireflux therapy.
This study included de-identified pH impedance tracings (Diversatek, Boulder, CO) from healthy volunteers and patients studied while on twice-daily (bid) PPI therapy, obtained from GERD centers in Europe and North America. Contributions of such studies were transmitted via secure web sites and downloaded for repeated analysis. We included data from 3 groups: 1) tracings from healthy volunteers recruited by the individual motility centers in Europe and North America by advertisement, who received bid PPI for 5–7 days; 2) tracings from heartburn-predominant proven GERD patients from Europe with prior abnormal reflux monitoring off PPI, studied on bid PPI, data from some of which have been previously published8; and 3) tracings from a previously published cohort of regurgitation predominant proven GERD patients with prior abnormal reflux monitoring off PPI, studied on bid PPI therapy (PPI arm of a prospective PPI and magnetic sphincter augmentation [MSA] study performed in North America).6 In both comparison cohorts, patients were studied on PPI regardless of symptom response status with bid PPI, and therefore these cohorts included both responders and nonresponders to PPI therapy. This post hoc analysis involved deidentified pH impedance tracings with no links to the original study subjects, so institutional review board approval was not deemed to be necessary.
Ambulatory pH Impedance Tracings
Study contributors attested to standardized methodology used in each instance. All studies were performed using Diversatek equipment (Boulder, CO) after at least 6 hours of fasting, and on bid PPI regimen. An experienced motility technician positioned the pH impedance catheter so that the pH sensor was located 5 cm above the proximal border of the manometrically located lower esophageal sphincter (LES). The configuration of the catheter allowed monitoring impedance values at 4 levels in the distal (3, 5, 7, and 9 cm above the LES) and 2 levels in the proximal esophagus (15 and 17 cm). A subset of the healthy volunteer studies used a bifurcated pH impedance catheter. Most of the remaining tracings included a gastric pH sensor. During the recording, subjects recorded meal start/end times and body positions (upright vs recumbent) and continued their usual daily activities. Absence of manometric diagnoses (ie, normal high-resolution manometry [HRM]) was required for inclusion to the original studies, but analysis of manometric studies was not part of this study protocol and manometric recordings were not solicited. The presence or absence of a hiatus hernia based on HRM and/or endoscopy was extracted from the original study databases, and hiatus hernia >3 cm was an exclusion criterion for the original studies.
Consensus pH Impedance Analysis
Healthy volunteer tracings were excluded if any symptoms were reported during the study. All healthy volunteer tracings, and tracings from the heartburn-predominant European cohort were analyzed by at least 2 expert reviewers among the authors (CPG, RT, FZ, SR, ES, DS) working together via video teleconference. Meal times were identified and excluded from the analysis. The Wingate Consensus criteria were used for identification of reflux episodes and PSPW, and agreement by at least 2 reviewers was required to be considered.9 Artifacts and meal/drink–induced pH drops were edited, but slow pH drifts without preceding swallows or impedance reflux were retained for calculation of AET. Mean nocturnal baseline impedance (MNBI) was averaged from the entire recumbent period, at channels 3 cm and 5 cm proximal to the LES, because this method (simplified MNBI) has been shown to correlate very well with the average baseline impedance from 3 discrete 10-minute periods in the recumbent position (conventional MNBI).10 The parameters included in the analysis were the following: total, upright, and supine AET; number of reflux episodes (total, acid, and nonacid); simplified MNBI at 3 and 5 cm above the LES; and PSPW index when available. Gastric pH data were included in the analysis when available.
European Comparison Cohort
Data from a subset of a cohort of patients from a previously published prospective study investigating factors associated with inadequate PPI response in an Italian cohort were also reviewed and compared with healthy volunteer data.8 This sample included only patients with heartburn, negative endoscopy, and abnormal AET on pH impedance monitoring while off PPI, who underwent a second pH impedance monitoring on bid PPI. This second pH impedance study was reviewed by study investigators using the consensus analysis methodology described above. Study reviewers were blinded to the symptom response to PPI therapy. When symptoms were reported in the GERD comparison cohorts, symptom index (abnormal when ⩾ 50%) and symptom association probability (abnormal when >95%) in relation to the dominant symptom were recorded; either of abnormal symptom index or abnormal symptom association probability constituted positive RSA. Summary data extracted after this analysis included: total, upright, and supine AET; numbers of reflux episodes; RSA; PSPW index; simplified MNBI; and gastric pH data.
North American Comparison Cohort
A second comparative cohort consisted of patients from the bid PPI arm of the prospective randomized controlled trial comparing magnetic sphincter augmentation and medical management of regurgitation-predominant GERD in North America.7 Only studies performed with the use of Diversatek equipment were included. Summary data extracted from this cohort included AET, numbers of reflux episodes, RSA, simplified MNBI, and gastric pH data when available, following sequential expert analysis by 2 reviewers (CPG, MFV).
Median values and interquartile ranges (IQRs) as well as 5th and 95th percentile values for AET, reflux episode numbers, simplified MNBI, and PSPW index were calculated for healthy volunteers and compared with corresponding values in the comparison cohorts. One-way analysis of variance with Tukey test, Mann-Whitney U test, Kruskal-Wallis followed by Dunn test, and chi-square test were used as appropriate. In all instances, P < .05 was required for statistical significance. To identify a clinically meaningful AET threshold capable of segregating symptom responders from nonresponders to determine when to escalate reflux therapy, the 2 comparison cohorts were analyzed with the use of receiver operating characteristic (ROC) curve analysis, both in combination and separately. Further analysis was performed using combinations of established nonphysiologic reflux metrics based on pH impedance monitoring off PPI (AET >4.0%, reflux episodes >80, MNBI <1500 ohms), as well as the combination of AET and reflux episode thresholds derived from ROC analysis. Performance characteristics of individual metrics and combinations of metrics are reported, including sensitivity, specificity, and area under the ROC curve (AUC). Improvement of ⩾ 50% in heartburn recorded on 4-point Likert-type scales (0 = no heartburn; 3 = severe heartburn) identified PPI responders as well as improvement following antireflux surgery (ARS) in the European comparison cohort. Similarly, ⩾ 50% improvement on GERD Health-Related Quality of Life scale identified PPI responders as well as improvement following subsequent MSA in the North American comparison cohort.
ResultsFrom the 77 pH impedance studies in healthy volunteers, 11 were excluded (symptoms reported during study = 8; supragastric belching = 2; technical artefact = 1), and 66 studies were analyzed (median age, 37.5 years; 29 female). Of 44 pH impedance studies on PPI in patients with heartburn from the European cohort, 1 was excluded because of dominant supragastric belching on the pH impedance study, and 43 studies were analyzed using similar methodology (median age, 57.0 years; 24 female). There were 42 proven GERD patients with regurgitation from the North American cohort (median age 41.6 years, 18 female). Healthy volunteers were significantly younger than both the North American and European comparison cohorts (P < .026 and P < .001, respectively) (Table 1). The European comparison cohort was older and had a higher proportion with hiatus hernia compared with the North American cohort (P ⩽ .01) (Table 1).
Table 1. Baseline Demographics and pH Impedance Metrics on Twice-Daily Proton Pump Inhibitor Therapy
AET, acid exposure time; MNBI, mean nocturnal baseline impedance; PSPW, post-reflux swallow-induced peristaltic wave.
a P < .001 compared with both comparison cohorts, except age vs North American comparison cohort, where P = .026.
b P ⩽ .01 compared with European comparison cohort.
c P =.013 compared with European comparison cohort.
Differences Between Healthy Volunteers and GERD Patients
Healthy volunteers on PPI had significantly lower AET and reflux episodes compared with GERD patients on PPI in both European and North American cohorts (P < .001 for each comparison) (Table 1). While median total AETs were similar (P = .09), median reflux episodes were higher in the North American cohort compared with the European cohort (66 and 40, respectively; P < .001). When Lyon Consensus thresholds were used, 7 patients (17.1%) of the European cohort and 3 (7.1%) of the North American cohort had distal AET >6%, a threshold that was not met by any of the healthy volunteers. In contrast, total reflux episodes were >80 in 17.1% of the European cohort and 28.6% of the North American cohort, and none of the healthy volunteers (Figure 1A). Within the healthy volunteers, 19 (28.8%) had PSPW index ⩾ 50% and 58 (87.9%) had PSPW index ⩾ 30%, compared with 3 (7.1%) and 13 (30.2%) of the European heartburn cohort (P = .006 for each comparison); PSPW was not available for the North American cohort. Percentage of time gastric pH <4.0 correlated only modestly with esophageal AET in healthy volunteers (n = 27; Pearson r = .59; P < .001); this correlation was poor in the European heartburn cohort (Pearson r = .31; P = .05), making this metric unhelpful.
Figure 1. (A) Comparisons of Lyon Consensus–based pH impedance thresholds for acid exposure time (AET) and reflux episode numbers. (B) Combinations of pH impedance metrics on pH impedance monitoring performed on bid therapy in healthy volunteers and heartburn- and regurgitation-predominant GERD patients. (C) Proportions of heartburn- and regurgitation-predominant nonresponders to bid proton pump inhibitor (PPI) therapy between categories of reflux evidence. Among individual metrics, proportions with abnormal AET were higher in heartburn-predominant presentations, and reflux episode numbers were higher in regurgitation-predominant presentations; both were significantly different from healthy volunteers (P < .001 for each comparison). Among combinations of metrics, conclusively abnormal metrics had similar proportions in heartburn- and regurgitation-predominant presentations, while borderline elevations were seen slightly more often in regurgitation; all were significantly different from healthy volunteers. While proportions of nonresponders were numerically higher in regurgitation-predominant patients, this reached statistical significance compared with heartburn predominant patients only in the borderline evidence category. There were more patients with heartburn among those with no reflux evidence.
Differences Between GERD Symptom Responders and Nonresponders to PPI Therapy
All patients had proven GERD and underwent pH impedance monitoring on bid PPI, regardless of their response to this therapy. Median AET and number of reflux episodes were not significantly different between responders and nonresponders to bid PPI in the 2 comparison cohorts (P ⩾ .1 for each comparison) (Table 2). Proximal reflux episodes trended higher in regurgitationpredominant nonresponders compared with responders (P = .09), but were similar between nonresponders and responders in heartburn-predominant patients (P = .48) (Table 2). Correlation between percentage of time gastric pH <4.0 and AET was similarly suboptimal both in responders (Pearson r = .24; P= .24) and nonresponders (Pearson r = 0.40; P = .11). There were no differences in simplified MNBI or PSPW index between responders and nonresponders in the European cohort, and no difference in simplified MNBI in the North American cohort. Within the European cohort, 12 (70.6%) of nonresponders and 18 (69.2%) of responders had PSPW index < 30% (P = .9), and none of nonresponders and 3 of responders had PSPW index <50% (P = .15). Lyon Consensus metrics were not discriminative of symptom response in either of the 2 comparison cohorts when used individually (Table 3).
Table 2. Comparison of Demographics and pH Impedance Metrics Between Proton Pump Inhibitor Responders and Nonresponders
AET, acid exposure time; MNBI, mean nocturnal baseline impedance; PPI, proton pump inhibitor; PSPW, post-reflux swallowinduced peristaltic wave.
Table 3. Breakdown of pH Impedance Metrics in Patients With Proven Gastroesophageal Reflux Disease Based on Response to Proton Pump Inhibitor Therapy, n (%)
a P = .006 compared to nonresponder.
Establishing Thresholds for AET and Reflux Episodes for “on PPI” pH Impedance Studies
When the 2 comparison cohorts were pooled for ROC analysis, an AET threshold of 0.5% predicted PPI nonresponse with sensitivity 0.62, specificity 0.51, AUC 0.58, and P = .22. Reflux episode threshold of 40 episodes performed better at predicting nonresponse, with sensitivity 0.80, specificity 0.51, AUC 0.70, and P = .002 (Figure 2).
Figure 2. Receiver operating characteristic (ROC) analysis demonstrating optimal thresholds for numbers of reflux episodes (40 episodes) and acid exposure time (AET; 0.5%) predicting symptom improvement in the pooled GERD patient cohort.
Composite pH Impedance Metrics in Interpreting pH Impedance Studies on bid PPI
Because our analysis demonstrated that single metrics do not reliably identify responders from nonresponders, further analyses were performed using multiple metrics sequentially (Figure 3). Patients were categorized first into those with abnormal reflux burden and abnormal mucosal integrity. This consisted of nonphysiologic AET (>4.0%),2,11 conclusively elevated reflux episodes (>80 episodes),2,6 and lower MNBI than that described in healthy asymptomatic volunteers (<1500 ohms).11 This category included 32.6% of the European heartburn cohort and 40.5% of the North American regurgitation cohort, but no healthy volunteers (Figure 1B). There were only 3 patients with PSPW index >30% (21.4%) and none with PSPW index >50% in this category. In addition, there were numerically more patients with regurgitation without 50% symptom response to PPI therapy (82.4%) than heartburn patients (57.1%) but without statistical significance (P = .12) (Figure 1C). Performance characteristics of the presence of abnormal reflux burden and/or abnormal mucosal integrity in predicting PPI nonresponse consisted of sensitivity 0.50, specificity 0.71 and AUC 0.59 (P = .15) (Table 4).
Figure 3. Schematic demonstrating reflux evidence categories with actual numbers of patients with each abnormality, and proposed management implications. AET, acid exposure time; MNBI, mean nocturnal baseline impedance; PSPW, post-reflux swallow-induced peristaltic wave; RSA, reflux-symptom association.
Table 4. Performance Characteristics of Individual and Composite Reflux Metrics From pH Impedance Monitoring “On PPI” in Predicting Persisting Symptoms Despite PPI Therapy
Using AET and reflux episode thresholds determined from the present study, patients with borderline reflux burden (AET >0.5% but <4%, 40–80 reflux episodes) were further identified (Figure 3). This included more regurgitation patients (21 patients, 50.0%) compared with heartburn patients (14 patients, 32.6%). Only 1 (7.1%) of the heartburn patients had PSPW index >50%, and 5 patients (37.5%) had PSPW index >30%. There was a high proportion of regurgitation patients (81.0%) without symptom response in this category, compared with heartburn patients (28.6%; P = .002). Performance characteristics of abnormal and borderline reflux burden categories together in predicting PPI nonresponse consisted of sensitivity 0.86, specificity 0.36, and AUC 0.62 (P = .07).
Proportions with reflux hypersensitivity (RSA with otherwise normal metrics) were low. Finally, more heartburn patients had a normal study (32.5%) compared with regurgitation patients (4.8%). While there were more heartburn patients with symptom response (71.4%) in the category without reflux evidence, the numbers of patients were too small to make meaningful comparisons in both of these categories. Among the European heartburn cohort with no reflux evidence, 5/14 (35.7%) had PSPW >30% and 2/14 (14.3%) had PSPW >50%.
Outcome Following Escalation of GERD Management Based on pH Impedance Findings in PPI Nonresponders
Within the subset of patients with abnormal reflux burden and/or abnormal mucosal integrity when tested “on PPI,” 20 patients underwent invasive GERD management, of which 5 underwent ARS (European cohort) and 15 underwent MSA (North American cohort). At follow-up of at least 6 months, >50% symptom reduction was achieved by 17 patients (85.0%), consisting of 3/5 ARS patients and 14/15 MSA patients. Within responders, 6 (35.3%) had a hiatus hernia: 100% of ARS responders and 21.4% of MSA responders. Within nonresponders, 2 (66.7%) had a hiatus hernia, 1 each in the ARS and MSA nonresponders.
Within the subset with borderline reflux burden, 9 patients underwent invasive GERD management, all of whom underwent MSA (North American cohort), and 6 (66.7%) had ⩾ 50% symptom reduction at 6 months. Among these 9 patients, 2 had a hiatus hernia, and both responded to MSA. In sum, 29 patients with abnormal or borderline reflux burden underwent invasive GERD management, and 23 (79.3%) had ⩾ 50% symptom improvement, of which 8 (34.8%) had a hiatus hernia.
DiscussionIn this cohort study aimed to characterize reflux metrics on pH impedance monitoring performed in proven GERD patients on bid PPI therapy, we demonstrate that combinations of metrics have better value than individual metrics in first identifying symptomatic nonresponders to PPI and subsequently predicting response to escalation of antireflux management. Higher proportions of PPI nonresponders were identified among patients with combinations of abnormal reflux burden (AET >4%, >80 reflux episodes, and/or MNBI <1500 ohms), of whom 85% improved when invasive GERD management (ARS or MSA) was performed. On ROC analysis, >40 reflux episodes had better performance characteristics than AET >0.5% in predicting PPI nonresponders. When patients fulfilling these ROC-based reflux metrics (40–80 reflux episodes and/or AET 0.5%-4%) were pooled with those with abnormal reflux burden, 79.3% had ⩾ 50% symptom improvement with ARS or MSA. We conclude that pH impedance monitoring while on bid PPI identifies combinations of abnormal metrics that might direct escalation of GERD management to ARS or MSA in symptomatic patients, but further prospective outcome studies are needed to better refine thresholds.
The concept of performing pH impedance studies while on PPI in symptomatic patients was first introduced in the early 2000s,12 but the recommendation of using this modality only in proven GERD was reinforced by the international GERD consensus within the past 5 years.1,2 Our findings indicate that interpretation of pH impedance studies on bid PPI is not straightforward and requires additional information on the patient’s clinical profile,13 starting with prior evidence demonstrating pathologic GERD off PPI and the dominant symptom that remains unresolved with bid PPI therapy. The interpretation itself cannot hinge on a single metric applicable to all patients; instead, a plurality of abnormal metrics carries more weight than any single metric (Figure 3), unlike pH monitoring while off PPI, where abnormal AET is a strong single metric defining pathologic GERD and predicting treatment outcome.14 Despite these caveats, pH impedance monitoring while on bid PPI provides information that neither wireless pH monitoring, catheter-based pH monitoring, or endoscopic mucosal impedance (mucosal integrity) can provide: identification of the association between symptoms and nonacid reflux, characterization of the gastric pH profile, recognition of persistently low MNBI, and exclusion of behavioral syndromes such as supragastric belching and rumination.
When reflux monitoring studies are performed off PPI therapy, distal AET >4% is consistently designated as elevated,15–17 either borderline when 4%–6% or pathologic when >6%.2 Therefore, it is reasonable to assume that distal AET >4% would definitively be abnormal in studies performed on PPI therapy, and a threshold similar to this (4.2%) was successfully used in a randomized outcome study comparing medical and surgical management of refractory heartburn-predominant GERD.17 In addition, >80 reflux episodes, as proposed by the Lyon Consensus, has been demonstrated to predict improvement in regurgitation following invasive therapy with MSA.6 Finally, low MNBI is considered to be a longitudinal marker of reflux-induced mucosal damage, and a threshold of 1500 ohms was described as the lower threshold of normal (5th percentile value) in a large multicenter multinational normative cohort study.11 We identified subsets of both heartburn- and regurgitation-predominant patients fulfilling these characteristics in sequence, giving more weight to AET >4% and >80 reflux episodes, which accounted for 32.6% of heartburn patients and 40.5% of regurgitation patients in our study. We found that all patients with MNBI <1500 ohms were included within the former 2 parameters. As many as 57% of heartburn nonresponders and 81% of regurgitation nonresponders were included within this category, with sensitivity and specificity of 0.50 and 0.71, respectively, in predicting PPI nonresponders. Despite these modest performance characteristics, as many as 85% of the 20 patients who underwent escalation of GERD management to ARS or MSA improved symptomatically, suggesting that these parameters define a need for escalation of reflux management. The fact that previous studies indicate that patients with AET >4.2% improved to a greater extent with surgical management (62% improvement) compared with medical management alone partially supports this management direction.17
Our ROC analysis identified a reflux episode threshold of 40 in predicting nonresponse, with acceptable performance characteristics, especially in regurgitation-predominant patients. In contrast, an AET threshold of 0.5% had modest performance in predicting nonresponse, which is lower than the 1% threshold reported by Charbel et al with the use of pH monitoring alone in a retrospective cohort.5 We included these 2 parameters in hierarchic fashion in describing a borderline category of patients when evidence of abnormal reflux burden or abnormal mucosal integrity was not identified, which accounted for 32.5% of heartburn and 50% of regurgitation patients. Within this category, nonresponse was identified in 26% of heartburn and 80% of regurgitation patients, indicating a difference in response to the 2 cardinal reflux symptoms. We interpret this finding to indicate a need for alternate documentation of abnormal GERD pathophysiology, perhaps in the form of a disrupted esophagogastric junction (EGJ) (decreased EGJ tone vs hiatus hernia) before further escalation in management. When this borderline category was combined with the abnormal reflux burden category, sensitivity in predicting persisting symptoms despite PPI increased to 0.86, and as many as 79% of 29 patients undergoing ARS or MSA improved, indicating that our ROC-based AET and reflux episode thresholds may have value as markers for escalation of GERD management.
Of the remaining patients, reflux hypersensitivity (manifest as RSA without abnormal or borderline reflux burden) accounted for a small minority in both heartburn and regurgitation predominant cohorts, where management might include neuromodulators and complementary approaches such as diaphragmatic breathing or cognitive/behavioral therapy.18–20 Finally, when no pH impedance parameters were abnormal, >70% of heartburn patients were designated as responders. There were only 2 regurgitation patients who fit this category, further highlighting differences between heartburn- and regurgitationpredominant patients when pH impedance studies are performed while on PPI therapy. These findings are consistent with existing literature indicating suboptimal response of regurgitation with PPI therapy,21 in contrast to heartburn despite the fact that both heartburn (a perceptive symptom) and regurgitation (a transit symptom) are considered to be typical reflux symptoms.22 We demonstrate clear differences in reflux burden and symptom response between heartburn-predominant and regurgitation-predominant GERD patients, indicating that identification of the dominant nonresponding symptom is an important aspect of further evaluation and management of the refractory GERD patient (Figure 3). This is particularly relevant because regurgitation as a symptom in the context of proven GERD improves better with invasive approaches, including ARS,23 transoral incisionless fundoplication, 24 and MSA,25,26 compared with bid PPI therapy.
We report deviations in our study from previous literature regarding gastric pH measurements, MNBI, and PSPW index. Although PPI efficacy and therapeutic performance has been measured using duration of suppression of gastric pH in research settings,27,28 we were not able to demonstrate a consistent relationship between gastric pH and esophageal AET in GERD patients on bid PPI. Thus, gastric pH does not perform as well in clinical practice as in research settings, potentially related to variation in PPI efficacy (as multiple PPIs were used by the subjects in this study), variation in diet, and variation in location of the pH sensor in the stomach. Newer pH impedance metrics were assessed, but MNBI values were similar between PPI responders and nonresponders, which is consistent with existing data demonstrating normalization of esophageal mucosal integrity with PPI therapy.29 Although the PSPW index was lower in GERD patients compared with healthy volunteers, we could not demonstrate a difference in median PSPW index between responders and nonresponders, and proportions with PSPW index <50% and <30% were also similar, potentially related to small sample size. Studies performed in healthy volunteers have demonstrated wide variations in PSPW,12 suggesting that even a low PSPW index is compatible with the asymptomatic state or PPI response. However, proportions with PSPW index >30% was higher in those with no reflux evidence (35.7%) compared with abnormal reflux burden (21.4%).
Our study has several strengths. This is the first study that characterizes the value of pH impedance monitoring in proven GERD patients on bid PPI, compares findings with healthy volunteers on bid PPI, and provides guidelines for when to escalate antireflux therapy. The healthy volunteers and GERD patients were prospectively collected, and the latter were part of prospective studies where all-comers were studied with pH impedance monitoring while on bid PPI. The GERD patient cohorts included all-comers, ie, patients were studied with pH impedance studies on bid PPI regardless of symptom response with therapy, which was possible only because of prospective methodology used by the original studies from which the cohorts were extracted. We were able to ascertain that none of the patients had esophageal motility disorders that could have confounded esophageal clearance. The analysis of pH impedance studies were performed by at least 2 international experts in this multinational collaborative study. Using this strategy, clear differences were identified between studies performed on healthy volunteers and GERD patients, and between our analysis and existing literature on pH impedance parameters.
However, our study is not without limitations. The study represents retrospective analysis of prospectively collected data, without a uniform protocol for diet and type and dose of PPI. Our information regarding esophageal motor function and EGJ morphology was extracted from datasheets and original study inclusion/exclusion criteria rather than direct review of HRM or endoscopy studies. The validated questionnaires used in the 2 GERD cohorts were different as the original intention for symptom assessment varied between the studies. We also did not have body mass index values for comparison, although we were able to determine based on original study criteria that none of the patients in either the European or North American cohort had body mass index >35 kg/m2. Because of small sample sizes and use of ROC curves, there is risk for overfitting with limited degrees of freedom, which we acknowledge; a power calculation was not possible as we used all patients that could be included. While we have proposed threshold criteria to decide need for escalation of antireflux therapy, our outcome data are limited to patients who underwent ARS or MSA rather than for the entire cohort, and we have used historical outcome data to substantiate our recommendations. There were not enough patients with RSA without abnormal AET, and use of neuromodulators was not available for review in either of the 2 cohorts; it is quite possible that symptomatic patients without abnormal reflux burden could have benefited from symptomatic therapy including neuromodulators to supplement PPI therapy. Finally, consensus analysis between experts is difficult to perform or reproduce in clinical practice, although careful manual analysis of pH impedance studies can improve the quality of interpretation. Nevertheless, we feel our analysis provides helpful direction for clinicians evaluating symptomatic GERD patients despite bid PPI therapy.
In conclusion, combinations of abnormal pH impedance metrics on studies performed on bid PPI therapy in symptomatic patients with proven GERD can help determine the need for escalation of antireflux therapy. The presenting symptom is important in decision making, because there is more evidence to support escalation in regurgitation-predominant presentations. Our findings and thresholds will be useful in planning prospective outcome studies to conclusively determine when to escalate antireflux therapy when GERD symptoms persist despite bid PPI therapy.
Abbreviations used in this paper: AET, acid exposure time; ARS, antireflux surgery; AUC, area under the receiver operating characteristic curve; EGJ, esophagogastric junction; GERD, gastroesophageal reflux disease; HRM, high-resolution manometry; LES, lower esophageal sphincter; MNBI, mean nocturnal baseline impedance; MSA, magnetic sphincter augmentation; PPI, proton pump inhibitor; PSPW, post-reflux swallow-induced peristaltic wave; RSA, reflux-symptom association; ROC, receiver operating characteristic.
Address correspondence to: C. Prakash Gyawali, MD, Division of Gastroenterology, 660 South Euclid Ave, Campus Box 8124, Saint Louis, Missouri 63110. e-mail: cprakash@ wustl.edu; fax: (314) 454-5107.
CRediT Authorship Contributions
C. Prakash Gyawali, MD, MRCP (Conceptualization: Equal; Formal analysis: Equal; Investigation: Equal; Methodology: Lead; Writing – original draft: Lead; Writing – review & editing: Equal); Radu Tutuian, MD (Data curation: Equal; Investigation: Equal; Writing – review & editing: Supporting); Frank Zerbib, MD, PhD (Investigation: Supporting; Methodology: Supporting; Resources: Supporting; Writing – review & editing: Supporting); Benjamin Dale Rogers, MD (Formal analysis: Lead; Investigation: Supporting; Writing – original draft: Supporting; Writing – review & editing: Supporting); Marzio Frazzoni, MD, PhD (Resources: Supporting; Writing – review & editing: Supporting); Sabine Roman, MD, PhD (Investigation: Supporting; Methodology: Equal; Resources: Supporting; Writing – review & editing: Supporting); Edoardo Savarino, MD, PhD (Investigation: Supporting; Methodology: Supporting; Resources: Supporting; Writing – review & editing: Supporting); Nicola de Bortoli, MD, PhD (Resources: Supporting; Writing – review & editing: Supporting); Marcelo F. Vela, MD (Investigation: Supporting; Writing – review & editing: Supporting); Daniel Sifrim, MD, PhD (Conceptualization: Equal; Investigation: Equal; Methodology: Equal; Writing – original draft: Equal; Writing – review & editing: Equal).
Conflicts of interest
CPG: consultant for Medtronic, Diversatek, Ironwood, Isothrive, Takeda, and Quintiles. FZ: consultant/speaker for Takeda, Allergan, Biocodex, Vifor Pharma, Mayoli Spindler, Ipsen, Abbott, Reckitt Benckiser, Alfasigma. SR: consulting for Medtronic, research support from Diversatek Healthcare, Medtronic. ES: lecture fee from Medtronic, Takeda, Janssen, MSD, Abbvie, Malesci; consulting for Medtronic, Takeda, Janssen, MSD, Reckitt Bencikser, Sofar, Unifarco, SILA, Oftagest. NdB: speaker for Malesci. MFV: consultant for Medtronic, research support from Diversatek; DS: research grants from Reckitt Benckiser UK, Jinshan Technology China, Alfa Sigma Italy. All other authors declare no conflicts.
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.