Impaired Esophageal Mucosal Integrity and the Related Motility Factors in Refractory Gastroesophageal Reux Disease

Backgrounds. Research on esophageal mucosal integrity in gastro-oesophageal reux disease (GERD) has been taken seriously in recent years, especially in refractory GERD. Mean nocturnal baseline impedance (MNBI) is proposed as an indicator of impaired mucosal integrity. We aimed to compare the MNBI value and investigate the impact of esophageal motility on MNBI in different subtypes of refractory GERD. Methods. Endoscopy, high-resolution manometry, and 24h impedance-pH monitoring were performed in 412 refractory GERD patients. From these patients, 94 erosive esophagitis (EE), 52 non-erosive reux disease (NERD), and 31 Barrett's esophagus (BE) patients were enrolled in the study. EE group 54 included Los Angeles (LA) Grade-A/B and 40 LA Grade-C/D patients. 52 functional heartburn (FH) patients were used as the control. MNBI was acquired at 3 and 5 cm above lower esophageal sphincter (LES) and was compared between groups. Parameters of esophagogastric junction (EGJ) and LES, along with esophageal peristaltic sequences were recorded. Univariate and multivariate regression analysis were performed to determine the impact of these motility factors on MNBI in different subtypes of refractory GERD. were signiantly lower in all subgroups of refractory GERD patients than in FH patients. MNBI in NERD patients was similar with LA-A/B and LA-C/D patients. MNBI in NERD and LA-C/D patients was signiantly lower than in BE patients. No difference in MNBI was found between LA-A/B and BE patients. Ineffective esophageal motility and absent contractility were the risk factor for decreased MNBI in LA-A/B and LA-C/D patients, respectively. Type III EGJ (hiatus hernias) and decreased LES length were the risk factor for decreased MNBI in NERD and BE patients, respectively. and above of LA-A/B, , and functional heartburn (FH) patients.

shown clinical e cacy in approximately 80% of EE patients [4]. Nevertheless, about half of NERD patients with daily PPIs treatments still complain of symptoms [5,6]. The symptoms, such as heartburn and regurgitation, which are not relieved after 8-12 weeks of double dose PPIs treatments, are called refractory GERD [7,8]. Re ux hypersensitivity, non-acid re ux, re ux clearance obstacles, and many other factors result in PPI failure, among which impaired mucosal integrity plays a signi cant role [9]. Poor response of PPI and and persistant symptoms were found to be associated with esophageal mucosal impairments and ultrastructural alteration in patients with refractory GERD [10][11][12].
The morphological and functional impairments of esophageal mucosal integrity, such as dilated intercellular space (DIS), basal cell hyperplasia, tight junction protein changes, and increased epithelial permeability in GERD patients, have been widely recognized. However, the detection of such alterations is di cult to be applied in clinic [13,14]. Multichannel intraluminal impedance-pH (MII-pH) monitoring is usually used to detect re ux episodes of GERD patients. When there is no re ux or swallowing attack, the esophageal wall will contact the MII-pH sensor catheter directly, and the impedance at this time re ects the inherent conductivity of esophageal wall, that is, the baseline impedance (BI). The decreased value of BI re ects the impairment of mucosal integrity caused by re ux [15,16]. Considering that frequent re ux and swallowing activities disturb the measurement during daytime, mean nocturnal baseline impedance (MNBI), the average BI value of three 10 minutes' measurement at night (around 1:00, 2:00, and 3:00 AM) was then applied [17]. This retrospective study intended to investigate the characteristic of impaired esophageal mucosal integrity represented by decreased MNBI and the respective motility factors in refractory EE, NERD and BE patients based on high-resolution esophageal manometry (HREM) and MII-pH monitoring, in order to provide ideas for optimizing the treatment of refractory GERD. 94 EE patients were grouped into Los Angeles Grade A or B (LA-A/B, n = 54) and Grade C or D (LA-C/D, n = 40). 52 functional heartburn (FH) patients who were matched for age, sex, height, and body mass index (BMI) with refractory GERD patients and had no motility disorders were selected for the control group.

Diagnostic Criteria
Patients with endoscopic mucosal breaks were diagnosed with EE. Patients with normal endoscopy and abnormal acid exposure were diagnosed with NERD. BE patients were con rmed by the columnar metaplasia of distal esophagus, whether intestinal type or not [18]. FH patients were con rmed by normal endoscopic manifestations, normal acid exposure, and negative symptom association [19]. It was reported that there was no difference in BI values, intercellular spaces and microscopic histological lesions between FH patients and healthy people [17,20,21]. FH patients were therefore used as the control group, and the decreased MNBI in refractory GERD patients compared with FH patients (p < 0.05) was supposed to represent the impaired mucosal integrity in this study.

Endoscopic Examination
According to the Los Angeles classi cation, EE was divided into four grades based on the degree of esophageal mucosal breaks through an endoscopy review [22]. Grade A: 1 or more mucosal breaks con ned to the mucosal folds and no longer than 5 mm in length. Grade B: 1 or more mucosal breaks more than 5 mm in length, but not continuous between the two mucosal folds. Grade C: continuous mucosal breaks no more than 75% of the esophageal circumference. Grade D: continuous mucosal breaks more than 75% of the esophageal circumference.
BE was diagnosed by endoscopy with pathological biopsies. The salmon color columnar epithelium, which displaced the pink squamous epithelium ≥ 1 cm above esophagogastric junction (EGJ) was distinguished under endoscopy, then 2 biopsy samples were taken below the squamous-columnar junction from each patient. Most patients were short-segment BE and showed tongue like projection instead of circumferential manifestation in our study, therefore biopsies were not carried out as the Seattle protocol of four-quadrant biopsy [23,24].

HREM Measurement
HREM with a 24-channel water-perfused catheter of 4.0 mm in diameter was utilized as previously described [25]. Patients were asked to fast foods for 12 h and water for 6 hours before the examination.
Drugs interfering esophageal motility were suspended temporarily. Patients were asked to stay in a supine position and underwent 10 swallows of 5 ml water with a 30 seconds interval under ambient temperature [17]. Esophageal motility parameters were analyzed by the computer software (MedView360 Contraction vigor of esophageal body was evaluated by distal contractile integral (DCI). Swallows were de ned as normal (DCI > 450 mmHg•cm•s and < 8000 mmHg•cm•s), weak (DCI > 100 mmHg•cm•s and < 450 mmHg•cm•s), or failed (DCI < 100 mmHg•cm•s), and a swallow with a large break (length > 5 cm) in the 20mmHg isobaric contour and DCI > 450 mmHg•cm•s was de ned as fragmented. Weak, failed and fragmented swallows were all de ned as ineffective. The diagnosis of ineffective esophageal motility (IEM) is de ned as ≥ 70% ineffective swallows or ≥ 50% failed swallows. Absent contractility was de ned as normal median integrated relaxation pressure (IRP) and 100% failed swallows. The morphology of EGJ depends on the relationship between lower esophageal sphincter (LES) and crural diaphragm (CD). Type I EGJ had overlapped LES and CD without separation, type II EGJ had a 1-2 cm separation between LES and CD, and type III EGJ had a > 2 cm separation between LES and CD. Type III EGJ was consistent with radiographically evident hiatus hernia. The vigor of EGJ barrier was measured over 3 consecutive respiratory cycles above the gastric pressure with a method analogous to that for evaluating DCI. It was then divided by the duration of 3 respiratory cycles to yield the EGJ contractile integral (EGJ-CI) independent of time. EGJ-CI of less than 25 mmHg•cm was considered hypotensive [26]. IEM, absent contractility, EGJ morphology, EGJ-CI, LES length, LES resting pressure, and IRP were recorded.

MII-pH Monitoring
MII-pH monitoring was performed off acid inhibitor for at least 7 days. HREM was used to locate LES. pH sensors of MII-pH catheter were positioned 5 cm above the LES, and impedance electrodes were positioned at 3, 5, 7, 9, 15, and 17 cm above the LES. Distal MNBI was measured at both 5 cm and 3cm above LES [27,28,29]. Esophageal re ux parameters were analyzed by the computer software (JSG.Zp.PlatformUI, Jinshan, China).
Acid exposure was de ned as abnormal if acid exposure time (AET) was beyond 6% and normal if AET was under 4%. When AET was between 4% and 6%, re ux events were used as an adjunctive metric, which was abnormal when > 80 per 24h and normal when < 40 per 24h [2]. Symptom association was determined as positive when symptom association probability (SAP) ≥ 95% or symptom Index (SI) ≥ 50%, otherwise negative [19]. AET, re ux events, SAP, and SI were recorded.

Statistical Analysis
Categorical variables are presented as frequencies and proportions, while continuous variables are expressed as mean ± standard error (normal distribution) or as median with interquartile ranges (nonnormal distribution). Chi-square tests were used for categorical variables. Kruskal-Wallis tests were used for continuous variables. Between-group comparison was performed using T test or Mann-Whitney U test.
Univariate and multivariate liner regression analysis was performed to identify motility related risk factors for impaired mucosal integrity, in which MNBI was taken as outcome variables and motility parameters were taken as dependent variables. Differences were con rmed as statistically signi cant when P < 0.05.
The statistical analysis was performed with SPSS software (version 20.0; IBM Corp, Armonk, NY, USA).

Clinical Characteristics of Patients
There was no statistical difference in age, sex, height, and BMI in the 5 groups (all P > 0.05). The proportion of IEM, absent contractility, type III EGJ, and hypotensive EGJ-CI ( Figure.1) was higher in refractory GERD patients than in FH patients (all P < 0.05). Differences in LES resting pressure and IRP were also found in the 5 groups (both P < 0.05) ( Table.1 Figure.2-b). Compared to FH patients, all four groups of GERD patients had signi cantly decreased MNBI values at both 5 and 3 cm above LES (all P < 0.05). MNBI values in LA-C/D patients were lower than in LA-A/B patients at 5 cm above LES and lower than in BE patients at both 5 and 3 cm above LES (all P < 0.05). MNBI values in NERD patients were lower than in BE patients at 3 cm above LES (P < 0.05) and there was a trend at 5cm above LES (P = 0.053). No difference in MNBI values was found between LA-A/B and BE, between LA-A/B and NERD, or between LA-C/D and NERD (all P > 0.05).

Linear Regression Analysis of MNBI in GERD
A signi cant decrease in MNBI values was found in all refractory GERD patients, then stepwise linear regression models were created to distinguish the motility factors related to mucosal impairments in each subgroup of refractory GERD patients. In LA-A/B patients, the presence of IEM (from 0 to 1) resulted in a signi cant decrease of MNBI values (β=-677.7, P = 0.024 at 5cm andβ=-628.6, P = 0.030 at 3cm) in both univariate and multivariate analysis (Table.2). In LA-C/D patients, IEM, absent contractility, and hypotensive EGJ-CI were signi cant when calculated in respective univariate analysis (all P < 0.05), however, only absent contractility (P = 0.007 at 5cm and P = 0.002 at 3cm) showed signi cance when all 3 metrics were included in multivariate analysis. The presence of absent contractility (from 0 to 1) resulted in a signi cant decrease of MNBI values (β=-658.3 at 5cm and β=-766.1 at 3cm) (Table.2).
In NERD patients, the presence of type III EGJ (from 0 to 1) led to a decrease of MNBI values (β=-319.1, P = 0.014 at 5cm) (Table.3). In BE patients, LES length and IRP were signi cant when calculated in respective univariate analysis (all P < 0.05), and only LES length showed signi cance (β = 644.2, P = 0.015 at 5cm and β = 667.2, P = 0.005 at 3cm) in multivariate analysis (Table.3). In other words, IEM and absent contractility were the risk factors for impaired mucosal integrity in LA-A/B and LA-C/D patients, respectively, in addition, type III EGJ and decreased LES length were the risk factors for impaired mucosal integrity in NERD and BE patients, respectively.

Discussion
This observational study aimed at comparing esophageal mucosal integrity in different subtypes of refractory GERD patients, including mild and severe EE, NERD and BE patients, as well as investigating the impact of abnormal anti-re ux barrier and esophageal body dysmotility on the mucosal integrity for the rst time. Compared with FH patients, signi cantly decreased MNBI values were found in refractory EE, NERD and BE patients. However, the severity of macroscopic mucosal changes was not consistent with the degree of microscopic impairments re ected by baseline impedance in the distal esophagus.
Furthermore, weakened esophageal body peristalsis, type III EGJ, and shortened LES were the risk factors for the impaired mucosal integrity in EE, NERD, and BE patients, respectively. We therefore conclude that HREM and MII-pH parameters were of great value in the identi cation and prediction of esophageal mucosal impairments in refractory GERD patients.
DIS and reduced transepithelial resistance (TEER) could morphologically and functionally distinguish the impaired mucosal integrity in GERD patients [30][31][32]. Earlier in an acid perfusion experiment, BI values decreased along with DIS and correlated with esophageal TEER positively [15]. Furthermore, BI values were found to be lower in patients with pathological acid exposure than in patients with physiological acid exposure, and both were lower than in healthy people [16]. Subsequently, BI was proposed an indicator of esophageal microscopic abnormalities caused by re ux and a marker of re ux burden [14][15][16]33]. FH refers to a burning sensation in the substernal region without endoscopic injury, abnormal acid exposure, and correlation between re ux events and symptoms [19]. The severity of histological impairments and the level of BI values in FH patients were comparable to those in healthy people [17,20,21], thus they were used as control in this study.
Decreased MNBI values to varying degrees were discovered in all groups of refractory GERD in our study. MNBI values of EE patients and NERD patients ranged from differences to no difference in previous studies [34][35][36]. No discrepancy in MNBI values between EE (both LA-A/B and LA-C/D) and NERD patients was found here, signifying a microscopic mucosal injury in NERD patients although no macroscopic mucosal injury at routine endoscopy. This was supported by the fact that DIS existed in NERD patients [15]. In addition, MNBI values were lower in BE patients than in control subjects, in accordance with Hemmink's research [37]. This result suggested that the esophageal region beyond Barrett's mucosa was also likely to be impaired, as MNBI was measured above Barrett's mucosa. Further research is needed to con rm the observation.
The diagnostic criteria for IEM in CCv4.0 is more strict than in CCv3.0, which contains larger proportions of ineffective or failed swallows, and fragmented peristalsis is also included in IEM in CCv4.0. In Reddy's study, the number of total re ux events (22.7% vs 9.0%, P < 0.01) and the bolus exposure time (3.4% vs 2.1%, P < 0.01) in GERD patients with more than 50% ineffective swallows were much more than those with normal motility. In addition, the total re ux events and AET increased with the increased proportion of ineffective swallows [38]. Rengarajan et al. reported that more severe dysmotility such as absent contractility was associated with abnormal re ux burden more than less severe dysmotility such as IEM [39]. The incidence of esophageal mucosal impairments increased with the increasement of re ux burden. In an HREM test of GERD patients undergoing 10 swallows, abnormal MNBI was related to > 70% ineffective swallows (P = 0.046), tended to be related to > 70% fragmented swallows (P = 0.076) and was unrelated to ≥ 50% failed swallows (P = 0.580) [40]. Our results showed that IEM and absent contractility were the risk factor for mucosal impairments in mild and severe refractory EE patients, respectively. A decreased MNBI value was the result of abnormal re ux burden, esophageal body dysmotility and re ux clearance obstacles [29,41,42].
The structure and function of EGJ play an essential role in the anti-re ux barrier, and the primary mechanism is the high-pressure band between LES and CD, preventing the occurrence of re ux. It is believed that re ux burden increases with the separation between LES and CD, and the pathological morphology of EGJ (a hiatus hernia) is a predictor of abnormal acid exposure [39,43,44]. Consequently, it was not surprising that decreased MNBI was related with type III EGJ in NERD patients here. EGJ-CI was a new indicator of EGJ contractility and it decreased with the upgrade of EGJ morphology, EGJ contractility gradually weakened with the increase of the gap between LES and CD [25,43]. EGJ-CI could not predict MNBI in Rogers' study but was able to predict AET in Rengarajan's study [39,40,44]. Our data showed that EGJ-CI was associated with MNBI at 3cm above LES in LA-C/D patients only in univariate analysis, not in multivariate analysis. We concluded that EGJ-CI has a de nite in uence on mucosal integrity but not as signi cant as EGJ morphology.
Re ux usually occurs via different mechanisms based on the anatomical and physiological variation of EGJ and LES, such as low LES pressure and transient LES relaxations [41]. Decreased LES resting pressure has been reported to be a predictor of increased acid exposure [38,45]. IRP re ects the relaxation function of LES. Neither LES resting pressure nor IRP was signi cantly related to MNBI values in our study. However, we found that the decreased LES length led to decreased MNBI values in BE patients, supported by the views that shortened esophagus was related to more re ux events and higher DeMeester score [46,47].
Limitations could not be avoided and impeded the reliability of our research, the predominant of which may be the limited sample size, but a certain number of subjects were required in regression analysis. All refractory GERD patients were recruited from only one hospital, which may lead to potential selection biases. The prevalence of BE is low and most are short-segment BE in Asia area, and thus the application of Seattle protocol is limited [48]. Only 2 biopsy samples were obtained from each patient in our study and the diagnosis of BE was likely to be inaccurate. Moreover, instead of healthy people, FH patients, whose MNBI level was comparable to healthy people, were used as a control group. Even so, our ndings still had values in evaluating the characteristic of mucosal integrity impairments and in improving the therapeutic strategy for refractory GERD.

Conclusions
This study improved our understanding of the relationship between esophageal motility dysfunction and impaired mucosal integrity in refractory GERD. It showed that the impairment of esophageal mucosa in NERD patients did not differ from that of EE patients, although the degree of lesion under endoscopy differed remarkably. Hiatus hernias was likely the reason for mucosal integrity impairments in refractory NERD patients. Protective strategies on esophageal mucosa and explorations of hiatus hernias should be paid more attention to in refractory NERD patients. Abnormal esophageal body peristalsis and decreased LES length might play an important role in the impaired mucosal integrity in refractory EE and BE patients, respectively. The therapy focusing on peristaltic disorders and LES function should be considered for EE and BE patients when PPIs response is not good. Abbreviations GERD: gastro-oesophageal re ux disease; BI: baseline impedance; MNBI: mean nocturnal baseline impedance; EE: erosive esophagitis; NERD: non-erosive re ux disease; BE: Barrett's esophagus; FH: functional heartburn; LA: Los Angeles; LES: lower esophageal sphincter; EGJ: esophagogastric junction; PPIs: proton pump inhibitors; DIS: dilated intercellular space; HREM: high-resolution esophageal manometry; MII-pH: multichannel intraluminal impedance-pH; BMI: body mass index; CCv: Chicago Classi cation version; DCI: distal contractile integral; IEM: ineffective esophageal motility; IRP: integrated relaxation pressure; CD: crural diaphragm; EGJ-CI: EGJ contractile integral; AET: acid exposure time; SAP: symptom association probability; SI: symptom Index; TEER: transepithelial resistance.

Declarations Acknowledgments
We thank all the editors and reviewers for their helpful suggestions on improving the quality of our article.

Funding Statement
This work was not supported by any funding.

Availability of data and materials
The data used to support the ndings of this study are available from the corresponding author upon request.

Consent for publication
Not applicable.

Con icts of Interest
The authors do not have any competing interests.

Authors' Contributions
Yu Cheng collected and analyzed the data, and wrote the manuscript. Zhijun Duan designed the study and modi ed the manuscript. Dong Yang, Xiaoyu Sun, Lixia Wang, Zhifeng Zhang, Wan Huang, Xiaoling Geng, Liping Su, Ping Zhang conducted the experiment. All authors approved the nal version of the manuscript.

Ethics approval and consent to participate
The Ethics Committee of First A liated Hospital of Dalian Medical University Ethical approved the report of scienti c research project. It con rmed that project program and informed consent were reviewed.