performed using SPSS 20.0. RESULTS: Results are presented in Table 1. Correlations compar-ing the 10s-window and 3 respiratory cycles were excellent (r .0.86; p,0.001) and thus,the data using 3 respiratory cycles is presented to account for respiratory variability. Thecorrelation between LESPI and EGJCI was r=0.35 (p ,0.05) suggesting that these metricsmeasure different attributes of the EGJ. The correlation between EGJCI and LESPI withinspiratory pressure was 0.61 (p,0.001) and 0.46 (p,0.001), respectively. The correlationbetween EGJCI and LESPI with end-expiratory pressure was 0.60 (p ,0.001) and 0.39(p,0.001), respectively. CONCLUSION: HRM permits quantification of multiple facets ofthe EGJ pressure as it pertains to its function as an antireflux barrier. However, as a singlemetric, the EGJCI has a stronger correlation withmeasures of EGJ pressure and less respiratoryvariability compared to the LESPI and may be a better overall estimate of barrier function.These metrics will require validation with objective outcome measures before being appliedin clinical practice.HRM metrics describing the antireflux barrier

(A) referenced to intra-atmospheric pressure (G) referenced to intra-gastric pressure

Tu1852

Esophagoprotection Mediated by Exogenous and Endogenous Melatonin in anExperimental Model of Reflux EsophagitisPeter C. Konturek, Iwona Brzozowska, Aneta Targosz, Slawomir Kwiecien, StanislawKonturek, Tomasz Brzozowski

Reflux esophagitis is a common clinical entity in western countries with approximately 30%of population experiencing the symptoms at least once every month. The imbalance betweenthe protective and aggressive factors lead to inflammation and damage of the esophagealmucosa. We compared the effect of exogenous melatonin and melatonin derived endoge-nously from L-tryptophan with that of pantoprazole or ranitidine in acid reflux esophagitisby ligation of the rat pylorus and the limiting ridge between the forestomach and the corpus.Four hours after the induction of gastric reflux, an increase in mucosal lesions associatedwith edema of the submucosa with the infiltration of numerous neutrophils and the fall inesophageal blood flow (EBF) were observed. Both, melatonin and L-tryptophan or pantopra-zole significantly reduced the lesion index (LI) and raised the EBF. Pinealectomy whichsignificantly decreased plasma melatonin levels aggravated LI and these effects were reducedby melatonin and L-tryptophan. Luzindole, the MT2 receptor antagonist, abolished themelatonin- and L-tryptophan-induced reduction in LI and the rise in EBF. L-NNA andcapsaicin which augmented LI and decreased EBF, also significantly reduced melatonin-and L-tryptophan-induced protection and hyperemia, both restored with L-arginine andCGRP added to melatonin and L-tryptophan. Upregulation of IL-1 β and TNF-α mRNAsand plasma IL-1β and TNF-α levels were significantly attenuated by melatonin and L-tryptophan. We conclude that melatonin protects against acid reflux-induced damage viaactivation of MT2 receptors mediated by NO, the vasoactive neuropeptide CGRP releasedfrom sensory nerves and the suppression of expression and release of TNF- α and IL-1.

Tu1853

Unique High Resolution Manometry Esophagogastric Junction PressureFeatures Among PPI Non-Responders Subgrouped by pH-Impedance StudiesFrédéric Nicodème, Kern Khanna, Maria Pipa Muniz, Peter J. Kahrilas, John E.Pandolfino

BACKGROUND: The current role of manometry in refractory gastroesophageal reflux disease(GERD) is to exclude motor disorders as a cause of the continued symptoms. Paradoxically,there are no data to support that manometry is useful to diagnose GERD or to stratify diseaseseverity. This study investigated whether new esophageal pressure topography (EPT) metricsquantifying antireflux competence of the esophagogastric junction (EGJ) could distinguishamong PPI non-responders (PPINR) subgroups. METHODS: 72 asymptomatic controls and88 PPI PPINR patients were included. All subjects underwent HRM (standard protocol of10 5 ml swallows) and the PPINRs also underwent 24h pH-impedance. PPINRs were dividedinto 3 subgroups based on parameters indicative of abnormal reflux: a) No Criteria: acidexposure ,2%, ,48 reflux events (RE) /24h, and symptom index ,20%; b) All Criteria:acid exposure .4%, .73 RE /24h; and symptom index .50%; and c) Some Criteria:patients not fulfilling inclusion into the other groups. EPT metrics were analyzed using theManoView software (Given Imaging, Duluth GA): 1) Separation between lower esophagealsphincter and crural diaphragm (LES-CD); 2) Basal end-expiratory EGJ pressure (EGJPexp),measured at the midpoint between adjacent inspirations for 3 respiratory cycles; 3) Basalinspiratory EGJ pressure (EGJPinsp), the corresponding inspiratory peaks; 4) EGJ contractileindex (EGJCI), determined by enclosing the domain of the EGJ area encompassing these 3respiratory cycles with the DCI tool setting the isobaric contour 2 mmHg above the gastricpressure and dividing by its duration to make it time-independent. RESULTS: There was asignificant difference in LES-CD separation and EGJPexp between PPINR and controls(Table). All Criteria patients were significantly different than controls in all EPT parameters,whereas there was overlap with controls and patients with No Criteria or Some Criteria.EGJPinsp and EGJCI were the only parameters that were different amongst the 3 PPINR

S-863 AGA Abstracts

subgroups. CONCLUSIONS: PPINR patients subgrouped by pH-impedance measures ofpersistent reflux exhibited varied degrees of EGJ incompetence when evaluated using novelEPT metrics. The best discriminators among PPINR subgroups and controls patients weremetrics quantifying the contribution of the crural diaphragm contraction to EGJ pressure(EGJPinsp and EGJCI). These findings suggest that novel analysis paradigms of HRM withEPT may be useful in distinguishing patients with functional heartburn from those withrefractory GERD.Characteristics of normal volunteers and patients groups (Median [IQR])

* Different from Normal (Mann-Whitney) † Different between subgroups of patients withAll Criteria and other subgroups of patients

Tu1854

In Patients With NERD, It Is Easy for Acid Reflux to Extend to the ProximalEsophagus, Compared With Healthy Subjects and Patients With Mild RefluxEsophagitisHirohito Sano, Katsuhiko Iwakiri, Noriyuki Kawami, Yuriko Tanaka, Choitsu Sakamoto

Background: We have reported that refluxed acid tends to extend more easily towards theproximal esophagus, according to the severity of reflux esophagitis (RE). However, the wayrefluxed acid extends to this area in patients with non-erosive reflux disease (NERD) ispoorly understood. Furthermore, the mechanisms of acid reflux episodes in patients withNERD are not clear. Methods: 14 patients with NERD, 15 patients with L A-grade A or Band 15 healthy subjects underwent esophageal manometry with a 21-lumen perfused assem-bly, whichmonitored pressure in the pharynx, the upper esophageal sphincter, the esophagealbody, the LES and the proximal stomach. Esophageal pH was monitored at 2 and 7cmabove the LES and recordings were taken for 3 hrs in a sitting position after a meal comprising692 kcal and 33 % fat. Results: There was no difference in the basal LES pressure betweenthe 3 groups. In healthy subjects and patients with mild RE, more than 95 % of all acidreflux episodes were TLESRs and all acid reflux episodes in patients with NERD wereaccompanied by TLESRs. There was no difference in the rate of TLESRs between the 3groups nor was there any difference in the rate of acid reflux during TLESRs at 2cm abovethe LES. The rate of acid reflux during TLESRs at 7cm above the LES in the group withNERD (37.5% [35.2-46.0], median [interquartile range]), however was significantly greaterthan that of healthy subjects (11.8% [0-17.2], p ,0.0001) and in patients with mild RE(30.5% [15.2-40.0], p=0.0154) and as well, the rate of acid reflux during TLESRs at 7cmabove the LES in patients with mild RE was significantly (p=0.0041) greater than that ofhealthy subjects. Conclusion: In patients with NERD, it is easy for acid reflux to extend tothe proximal esophagus, comparedwith healthy subjects and patients withmild RE. Consider-ing that proximal reflux is a very important factor in the perception of reflux, it is clear thatin patients with NERD, there is a high possibility of these patients having a symptomaticreflux episode.

Tu1855

Expression of Cathepsin-G and Protease Activated Receptor-4 in Erosive andNon-Erosive Reflux Disease in Comparison With Functional Heartburn andControlsDaniela Friedrichs, Arne Kandulski, Thomas Wex, Doerthe Kuester, Peter Malfertheiner

Background: Protease-activated receptor-4 (PAR4) is activated by trypsin and cathepsin G(CatG), which is released during inflammation and essential for neutrophil function. MucosalCatG and CatG-mediated PAR4-activation have been found to be involved in human colonicinflammation before, and dose dependent effects of PAR4-activation on visceral pain havebeen reported from animal studies. Aim: To analyze the role of PAR4 and CatG in esophagealmucosa of patients with gastroesophageal reflux disease (GERD) in comparison with func-tional heartburn (FH) and healthy controls. Methods: 82 patients were included, 27 patientswith ERD (erosive reflux disease), 18 patients with NERD (non-erosive reflux disease), 14patients with FH and 25 subjects without any reflux symptoms. After pausing potential acidsuppressive medications for 3 weeks, all patients underwent upper endoscopy and pH-impedance (MII-pH). At endoscopy, biopsies were taken 3 cm above the gastroesophagealjunction. Histopathological changes were assessed according to updated Ismail Beigi criteria.Gene expression of CatG and β-actin was analyzed by quantitative RT-PCR, PAR4 by TaqManassay. ELISA was performed to assess the mucosal content of CatG. Results: Insignificantly,CatG gene expression was 5-fold higher in GERD when compared to controls. No differenceswere measured between FH and GERD. Protein amounts of CatG in mucosal biopsies werelow and did not reach the detection level of the test. In contrast, PAR4 gene expression wassignificantly elevated up to 10-fold in ERD, NERD and FH when compared to controls(p,0.0001). No differences were obtained between the subgroups. Histological evaluationdemonstrated characteristic changes in GERD. Dilated intercellular spaces and a histologicalsum score in NERD and ERD differed significantly when compared with FH and controls.Correlation analysis reveals a slight association of CatG gene expression with basal cell

AG

AA

bst

ract

s