Diagnostic performance of the endoscopic pressure study integrated system (EPSIS): a novel diagnostic tool for gastroesophageal reflux disease

• Abstract
• Introduction
• Methods
• Study design and patients
• EPSIS
• MII-pH
• Data collection
• Statistical analysis
• Results
• Study population synthesis
• EPSIS for diagnosis of acid reflux disorders
• Discussion
• References

Abstract

Background The endoscopic pressure study integrated system (EPSIS) is a prototypic system for monitoring intragastric pressure (IGP) fluctuations that result from opening of the cardia during gastric distension. The performance of EPSIS for the diagnosis of gastroesophageal reflux disease (GERD) was evaluated.

Methods A retrospective analysis was conducted of data prospectively collected over a 2-year period from 59 patients who underwent gastroscopy, EPSIS, and 24-hour pH monitoring. Using a dedicated electronic device and a through-the-scope catheter, maximum IGP (IGP_max) and IGP waveform pattern (uphill/flat) were recorded.

Results The optimal IGP_max cutoff was 18.7 mmHg. IGP_max < 18.7 mmHg (sensitivity 74.2 %, 95 % confidence interval [CI] 56.8 – 86.3; specificity 57.1 %, 95 %CI 39.1 – 73.5) and flat pattern (sensitivity 71.0 %, 95 %CI 53.4 – 83.9; specificity 82.1 %, 95 %CI 64.4 – 92.1) were associated with GERD. “Double” EPSIS positivity (IGP_max < 18.7 mmHg and flat pattern) provided maximum specificity (85.7 %, 95 %CI 68.5 – 94.3), whereas “any” EPSIS positivity (IGP_max < 18.7 mmHg or flat pattern) provided maximum sensitivity (80.6 %, 95 %CI 63.7 – 90.8). Maximum specificity and sensitivity for nonerosive reflux disease (NERD) was > 70 %. In multivariate analysis, “double” EPSIS positivity was the strongest predictor of GERD (odds ratio [OR] 16.05, 95 %CI 3.23 – 79.7) and NERD (OR 14.7, 95 %CI 2.37 – 90.8).

Conclusion EPSIS emerges as a reliable adjunct to routine gastroscopy for GERD diagnosis, and might prove helpful for the stratification and management of patients with reflux disorders.

Introduction

Gastroesophageal reflux disease (GERD), which comprises erosive esophagitis and nonerosive reflux disease (NERD), poses a great challenge to gastroenterologists owing to high prevalence and demanding management that occupies a fair amount of gastroenterology clinical practice worldwide [1] [2] [3]. Apart from symptoms and endoscopic findings, other modalities are utilized for the diagnosis of GERD, with 24-hour pH-monitoring with or without impedance being the gold standard [2] [3]. A prototypic diagnostic tool – the endoscopic pressure study integrated system (EPSIS) – has been developed to allow the monitoring of intragastric pressure (IGP) fluctuations that result from cardia opening induced by CO₂ insufflation and gastric distension during endoscopy ([Fig. 1]). Gastric distension and increase in IGP are also established mechanisms underlying the pathogenesis of GERD [4] [5] [6]. The aim of this study was to evaluate the performance of this novel diagnostic system.

Methods

Study design and patients

This was a retrospective analysis of data collected prospectively at Showa University Koto Toyosu Hospital, Tokyo, Japan, from April 2016 to July 2018. Consenting patients who had symptomatic GERD and an indication for diagnostic esophagogastroduodenoscopy (EGD) were included in the study. Typical GERD symptoms (e. g. heartburn, chest pain, regurgitation) were included. Some patients also complained of belching and dysphagia. However, patients with cough or globus alone were not included. Additional exclusion criteria were age < 18 years, pregnancy, prior upper gastrointestinal surgery, major esophageal motility disorders, Zollinger – Ellison syndrome, and nitrate/calcium blocker use.

All study participants underwent EGD, EPSIS, and 24-hour pH-impedance monitoring (MII-pH). In selected cases, histopathologic assessment and high resolution manometry were performed to rule out eosinophilic esophagitis and major motility disorders.

The study was approved by the Showa University Research Ethics Committee (IRB Registration Number 18T7030). Informed consent was obtained from all participants.

EPSIS

EPSIS was performed as follows. 1) Routine EGD using high definition endoscopes (GIF-Q260Z/GIF-Q290Z; Olympus Medical Systems Corp., Tokyo, Japan), CO₂, and intravenous propofol to minimize patient discomfort from gastric distension. 2) A through-the-scope catheter (PR-V235Q; Olympus Medical Systems Corp.) was inserted and the tip was placed at 2 cm from the esophagogastric junction in the retroflexed position, along the lesser curvature ([Fig. 2]). Extra care was taken to keep the endoscope in a neutral position allowing only a “passive” grasping and release of the endoscope so as not to exert any force at the cardia. The other end of the catheter was connected to the internal pressure measuring device (TR-W550, TR-TH08, AP-C35; Keyence, Osaka, Japan). 3) 50 mL of air was flushed through the catheter to remove liquid droplets. 4) CO₂ was continuously insufflated within the stomach in order to trigger cardia opening, and IGP parameters (maximum IGP and waveform pattern of IGP, uphill/flat) were recorded ([Fig. 3]). The IGP was measured for 30 seconds during insufflation. In case of retching, the process was repeated over an additional 30 seconds after the patient had settled. Two examiners checked the waveform and determined the pattern in all cases.

MII-pH

Following 1 week off proton-pump inhibitors and overnight fasting, a 2.13-mm MII-pH catheter (Sandhill Scientific, Highlands Ranch, Colorado, USA) was placed transnasally with the esophageal electrode 5 cm above the lower esophageal sphincter (LES) and its position was radiologically verified. After removal 24 hours later, data were collected and analyzed using dedicated software (Bioview Analysis; Sandhill Scientific). For acid reflux positivity, a DeMeester score ≥ 14.72 or an acid exposure time ≥ 4.2 % was applied [2] [3] [7].

Data collection

Patient (age, sex, body mass index), endoscopic (Hill and Los Angeles classifications), EPSIS (IGP), and MII-pH (acid exposure time, DeMeester score, symptom index, symptom association probability) data were retrieved [2] [3] [8].

Statistical analysis

Normality was assessed using the Shapiro–Wilk test. Quantitative variables were normally distributed and are summarized as mean and standard deviation (SD). Categorical variables are reported as frequency counts and percentages. Acid reflux on MII-pH was set as the reference criterion variable.

Sensitivity, specificity, likelihood ratios, and predictive values of EPSIS were calculated. The optimal IGP_max cutoff was estimated using the point that maximizes the Youden’s J statistic (highest test efficiency). EPSIS parameters were combined to either maximize specificity (“double” EPSIS positivity defined as IGP_max ≤ 18.7 mmHg and flat pattern) or sensitivity (“any” EPSIS positivity defined as IGP_max ≤ 18.7 mmHg or flat pattern).

The t test was used for quantitative variables and the chi-squared or Fisher’s exact test was used for categorical variables. Variables yielding P < 0.2 in univariate analysis were considered for multivariate logistic regression. To decrease the risk of type I error, only classic risk factors for GERD such as age, body mass index, sex, hiatal hernia, and esophagitis were tested. An “all possible equations” method was used to create the logistic models. Area under the receiver operating characteristic curve and P-Hosmer–Lemeshow test were used to assess discrimination and calibration. Models were internally validated using 10-fold cross-validation. All tests were two-tailed and P < 0.05 was considered significant.

Statistical analysis was performed using STATA 14.2 (StataCorp, College Station, Texas, USA). Assuming an 80 % expected sensitivity/specificity, α = 0.05, β = 0.2, and a desired ±15 % precision, the required total sample size was 57 patients.

Results

Study population synthesis

A total of 59 patients were included, of whom 18 had NERD and 13 had erosive esophagitis. The remaining 28 patients comprised a heterogeneous group with functional heartburn, reflux hypersensitivity, or no proven disorder. No patients with functional chest pain or supragastric belching were detected. Additional baseline characteristics are summarized in Table 1 s in the online-only Supplementary material.

EPSIS for diagnosis of acid reflux disorders

The optimal cutoff for IGP_max was 18.7 mmHg (efficiency 69.5 %, Youden’s J index 38.7) (Fig. 1 s). When examining the whole patient population, an EPSIS flat pattern (sensitivity 71.0 %, 95 % confidence interval [CI] 53.4 – 83.9; specificity 82.1 %, 95 %CI 64.4 – 92.1) achieved higher efficiency than IGP_max for the diagnosis of GERD (sensitivity 74.2 %, 95 %CI 56.8 – 86.3; specificity 57.1 %, 95 %CI 39.1 – 73.5). “Double” EPSIS positivity provided maximum specificity (85.7 %, 95 %CI 68.5 – 94.3) and “any” EPSIS positivity provided maximum sensitivity (80.6 %, 95 %CI 63.7 – 90.8). In patients without esophagitis (n = 44), the diagnostic accuracy of EPSIS for detecting NERD was lower, but maximum specificity and sensitivity remained over 70 %. A detailed analysis is provided in Table 2 s.

The association between EPSIS and symptoms was investigated and a significant link was found only for flat pattern and symptom index positivity (P  = 0.03). On multivariate analysis, compared with known risk factors for acid reflux, “double” EPSIS positivity was the strongest predictor for GERD (odds ratio [OR] 16.05, 95 %CI 3.23 – 79.7) and NERD (OR 14.7, 95 %CI 2.37 – 90.8) (Table 3 s, Fig. 2 s).

Discussion

The performance of a novel diagnostic tool – EPSIS – was evaluated in this study. Based on our findings, a flat pattern of IGP and an IGP_max under 18.7 mmHg could distinguish GERD/NERD from disorders with similar symptoms, such as functional heartburn, reflux hypersensitivity, or no proven disorder, with fair – good accuracy, with results being somewhat better for the flat pattern. When both EPSIS criteria were combined, the diagnostic performance was enhanced, allowing more accurate diagnosis of GERD and NERD. EPSIS positivity, as a novel predictor for GERD, outperformed known risk factors, including erosive esophagitis. Finally, a multivariate model showed that a diagnostic model comprising EPSIS, age, and the presence of esophagitis could further aid GERD diagnosis.

The association of GERD and NERD with a flat pattern on EPSIS is not surprising, as it indicates prolonged and persistent opening of the cardia, subsequent CO₂ escape by slow belching, and failure of IGP build up. Similarly, a lower IGP threshold for opening of the cardia would be consistent with a less “airtight” valve mechanism and reduced LES function. In patients with GERD, both phenomena may facilitate reflux of acidic gastric content into the esophagus [9] [10]. This content can be highly acidic, especially when originating from the so-called acid pocket, an area of unbuffered gastric acid in the proximal stomach [11]. The proximity of a “defective” cardia, as detected through EPSIS, with this acid pocket, facilitates more prominent acid reflux and exposure, both of which are strongly associated with erosive esophagitis and NERD [11] [12]. Taken together, these considerations may explain the predilection of EPSIS to detect erosive esophagitis and NERD over their functional counterparts [12] [13]. When both flat pattern of IGP and lower IGP_max were present, the correlation of EPSIS with GERD or NERD became stronger, even outweighing traditional predictors of acid reflux, for example endoscopically documented erosive esophagitis, which is found in 30 % of untreated and < 10 % of treated patients [2]. In a similar manner, symptoms are not good markers for the presence of GERD, a fact also depicted in the moderate performance of symptom association indices [2]. Conversely, the flat pattern of IGP showed good correlation with symptom index positivity in the current study, probably reflecting a more persistent dysfunction of the cardia, accompanied by a higher likelihood of the easily perceptible belching and heartburn symptoms that are produced by persistent gas/acid liquid reflux [2] [3] [10]. EPSIS may be considered as a tolerance test by excessively insufflating CO₂ into the stomach to assess potential LES valve mechanism.

At this point, however, certain study limitations must be acknowledged. No healthy individuals were included in this study and patients manifesting GERD symptoms theoretically have a greater pre-test probability of acid reflux. However, the distribution of disorders in this cohort reflects the heterogeneity that clinicians would face in a real-life clinical scenario of patients with symptomatic reflux [2] [3] [14]. Moreover, the endoscopic parameters assessed in our study contain an element of subjectivity. High resolution manometry was not performed in all study patients because manometric data would not be incorporated in final assessment of EPSIS diagnostic performance. Similarly, biopsies to rule out eosinophilic esophagitis were only taken, at the discretion of the endoscopists, when there was a clinical and/or endoscopic suspicion and were not taken routinely from all participants. This may have been insufficient to completely rule out eosinophilic esophagitis. The use of propofol may also represent a potential limitation, although available data on its effect on LES lack unanimity [15]. In addition, the study was performed in a small number of patients and thus statistical power was limited. However, we did identify EPSIS as an independent predictor of GERD and further investigation with a larger sample size is therefore warranted.

In conclusion, EPSIS can be used for diagnostic purposes in GERD, either as a single test or as part of a composite model. Owing to a tight link with acid-reflux disorders, its use for patient stratification and management seems quite appealing. Taking these into account, EPSIS emerges as a promising and reliable adjunct to routine EGD for the diagnosis of GERD.

Competing interests

Dr. Inoue is an advisor to Olympus Corporation and Top Corporation. He has also received educational grants from Takeda Pharmaceutical Company and Olympus Corporation. Dr. Manolakis is a Hellenic Society of Gastroenterology grant holder.

Acknowlegment

We thank Mr. Tomohiro Joya who provided insight and expertise in technical aspect of EPSIS development that greatly assisted the study.

• References

• 1 Dent J, El-Serag HB, Wallander MA. et al. Epidemiology of gastro-oesophageal reflux disease: a systematic review. Gut 2005; 54: 710-717
  CrossrefPubMedSearch in Google Scholar
• 2 Gyawali CP, Kahrilas PJ, Savarino E. et al. Modern diagnosis of GERD: the Lyon Consensus. Gut 2018; 67: 1351-1362
  CrossrefPubMedSearch in Google Scholar
• 3 Vaezi MF, Pandolfino JE, Vela MF. et al. White paper AGA: Optimal strategies to define and diagnose gastroesophageal reflux disease. Clin Gastroenterol Hepatol 2017; 15: 1162-1172
  CrossrefPubMedSearch in Google Scholar
• 4 Bredenoord AJ, Weusten BL, Timmer R. et al. Relationships between air swallowing, intragastric air, belching and gastro-oesophageal reflux. Neurogastroenterol Motil 2005; 17: 341-347
  CrossrefPubMedSearch in Google Scholar
• 5 Holloway RH, Kocyan P, Dent J. Provocation of transient lower esophageal sphincter relaxations by meals in patients with symptomatic gastroesophageal reflux. Dig Dis Sci 1991; 36: 1034-1039
  CrossrefPubMedSearch in Google Scholar
• 6 Holloway RH, Hongo M, Berger K. et al. Gastric distention: a mechanism for postprandial gastroesophageal reflux. Gastroenterology 1985; 89: 779-784
  CrossrefPubMedSearch in Google Scholar
• 7 Hirano I, Richter JE. Practice Parameters Committee of the American College of Gastroenterology. ACG practice guidelines: Esophageal reflux testing. Am J Gastroenterol 2007; 102: 668-685
  CrossrefPubMedSearch in Google Scholar
• 8 Hansdotter I, Björ O, Andreasson A. et al. Hill classification is superior to the axial length of a hiatal hernia for assessment of the mechanical anti-reflux barrier at the gastroesophageal junction. Endosc Int Open 2016; 4: E311-E317
  Thieme ConnectPubMedSearch in Google Scholar
• 9 Sifrim D, Holloway R. Transient lower esophageal sphincter relaxations: how many or how harmful?. Am J Gastroenterol 2001; 96: 2529-2532
  CrossrefPubMedSearch in Google Scholar
• 10 Sifrim D, Holloway R, Silny J. et al. Composition of the postprandial refluxate in patients with gastroesophageal reflux disease. Am J Gastroenterol 2001; 96: 647-655
  CrossrefPubMedSearch in Google Scholar
• 11 Kahrilas PJ, McColl K, Fox M. et al. The acid pocket: a target for treatment in reflux disease?. Am J Gastroenterol 2013; 108: 1058-1064
  CrossrefPubMedSearch in Google Scholar
• 12 Yamasaki T, O’Neil J, Fass R. Update on functional heartburn. Gastroenterol Hepatol (NY) 2017; 13: 725-734
  PubMedSearch in Google Scholar
• 13 Yamasaki T, Fass R. Reflux hypersensitivity: a new functional esophageal disorder. J Neurogastroenterol Motil 2017; 23: 495-503
  CrossrefPubMedSearch in Google Scholar
• 14 Pace F, Casini V, Pallotta S. Heterogeneity of endoscopy negative heartburn: epidemiology and natural history. World J Gastroenterol 2008; 14: 5233-5236
  CrossrefPubMedSearch in Google Scholar
• 15 de Leon A, Ahlstrand R, Thörn SE. et al. Effects of propofol on oesophageal sphincters: a study on young and elderly volunteers using high-resolution solid-state manometry. Eur J Anaesthesiol 2011; 28: 273-278
  CrossrefPubMedSearch in Google Scholar

Corresponding author

• References

• 1 Dent J, El-Serag HB, Wallander MA. et al. Epidemiology of gastro-oesophageal reflux disease: a systematic review. Gut 2005; 54: 710-717
  CrossrefPubMedSearch in Google Scholar
• 2 Gyawali CP, Kahrilas PJ, Savarino E. et al. Modern diagnosis of GERD: the Lyon Consensus. Gut 2018; 67: 1351-1362
  CrossrefPubMedSearch in Google Scholar
• 3 Vaezi MF, Pandolfino JE, Vela MF. et al. White paper AGA: Optimal strategies to define and diagnose gastroesophageal reflux disease. Clin Gastroenterol Hepatol 2017; 15: 1162-1172
  CrossrefPubMedSearch in Google Scholar
• 4 Bredenoord AJ, Weusten BL, Timmer R. et al. Relationships between air swallowing, intragastric air, belching and gastro-oesophageal reflux. Neurogastroenterol Motil 2005; 17: 341-347
  CrossrefPubMedSearch in Google Scholar
• 5 Holloway RH, Kocyan P, Dent J. Provocation of transient lower esophageal sphincter relaxations by meals in patients with symptomatic gastroesophageal reflux. Dig Dis Sci 1991; 36: 1034-1039
  CrossrefPubMedSearch in Google Scholar
• 6 Holloway RH, Hongo M, Berger K. et al. Gastric distention: a mechanism for postprandial gastroesophageal reflux. Gastroenterology 1985; 89: 779-784
  CrossrefPubMedSearch in Google Scholar
• 7 Hirano I, Richter JE. Practice Parameters Committee of the American College of Gastroenterology. ACG practice guidelines: Esophageal reflux testing. Am J Gastroenterol 2007; 102: 668-685
  CrossrefPubMedSearch in Google Scholar
• 8 Hansdotter I, Björ O, Andreasson A. et al. Hill classification is superior to the axial length of a hiatal hernia for assessment of the mechanical anti-reflux barrier at the gastroesophageal junction. Endosc Int Open 2016; 4: E311-E317
  Thieme ConnectPubMedSearch in Google Scholar
• 9 Sifrim D, Holloway R. Transient lower esophageal sphincter relaxations: how many or how harmful?. Am J Gastroenterol 2001; 96: 2529-2532
  CrossrefPubMedSearch in Google Scholar
• 10 Sifrim D, Holloway R, Silny J. et al. Composition of the postprandial refluxate in patients with gastroesophageal reflux disease. Am J Gastroenterol 2001; 96: 647-655
  CrossrefPubMedSearch in Google Scholar
• 11 Kahrilas PJ, McColl K, Fox M. et al. The acid pocket: a target for treatment in reflux disease?. Am J Gastroenterol 2013; 108: 1058-1064
  CrossrefPubMedSearch in Google Scholar
• 12 Yamasaki T, O’Neil J, Fass R. Update on functional heartburn. Gastroenterol Hepatol (NY) 2017; 13: 725-734
  PubMedSearch in Google Scholar
• 13 Yamasaki T, Fass R. Reflux hypersensitivity: a new functional esophageal disorder. J Neurogastroenterol Motil 2017; 23: 495-503
  CrossrefPubMedSearch in Google Scholar
• 14 Pace F, Casini V, Pallotta S. Heterogeneity of endoscopy negative heartburn: epidemiology and natural history. World J Gastroenterol 2008; 14: 5233-5236
  CrossrefPubMedSearch in Google Scholar
• 15 de Leon A, Ahlstrand R, Thörn SE. et al. Effects of propofol on oesophageal sphincters: a study on young and elderly volunteers using high-resolution solid-state manometry. Eur J Anaesthesiol 2011; 28: 273-278
  CrossrefPubMedSearch in Google Scholar

• Supplementary Material