Abnormal Structure and Function of the Esophagogastric Junction and Proximal Stomach in...

The American Journal of GASTROENTEROLOGY VOLUME 109 | MAY 2014 www.amjgastro.com

ORIGINAL CONTRIBUTIONS nature publishing group658 E

SO

PH

AG

US

see related editorial on page x

INTRODUCTION Th e esophagogastric junction (EGJ) is the key defense against

refl ux of gastric contents into the esophagus. It comprises the

lower esophageal sphincter (LES) and the crural diaphragm that,

together with the clasp and sling fi bers of the gastric cardia, form

an integrated sphincter complex. Th is refl ux barrier is able to

Abnormal Structure and Function of the Esophagogastric Junction and Proximal Stomach in Gastroesophageal Refl ux Disease Jelena Curcic , PhD 1 , 2 , Shammodip Roy , M.Tech 3 , Alexandra Schwizer , BM, BCh 2 , Elad Kaufman , MD 2 , Zsofi a Forras-Kaufman , MD 2 , Dieter Menne , PhD 4 , Geoffrey S. Hebbard , MD, PhD 5 , Reto Treier , PhD 1 , Peter Boesiger , PhD 1 , 6 , Andreas Steingoetter , PhD 1 , 2 , Michael Fried , MD 2 , 6 , Werner Schwizer , MD 2 , 6 , Anupam Pal , PhD 3 , and Mark Fox , MD 2 , 7

OBJECTIVES : This study applies concurrent magnetic resonance imaging (MRI) and high-resolution manometry (HRM) to test the hypothesis that structural factors involved in refl ux protection, in particular, the acute insertion angle of the esophagus into the stomach, are impaired in gastroesophageal refl ux disease (GERD) patients.

METHODS: A total of 24 healthy volunteers and 24 patients with mild-moderate GERD ingested a test meal. Three-dimensional models of the esophagogastric junction (EGJ) were reconstructed from MRI images. Measurements of the esophagogastric insertion angle, gastric orientation, and volume change were obtained. Esophageal function was assessed by HRM. Number of refl ux events and EGJ opening during refl ux events were assessed by HRM and cine-MRI. Statistical analysis applied mixed-effects modeling.

RESULTS: The esophagogastric insertion angle was wider in GERD patients than in healthy subjects ( + 7 ° ± 3 ° ; P = 0.03). EGJ opening during refl ux events was greater in GERD patients than in healthy subjects (19.3 mm vs. 16.8 mm; P = 0.04). The position of insertion and gastric orientation within the abdo-men were also altered (both P < 0.05). Median number of refl ux events was 3 (95 % CI: 2.5 – 4.6) in GERD and 2 (95 % CI: 1.8 – 3.3) in healthy subjects ( P = 0.09). Lower esophageal sphincter (LES) pressure was lower ( − 11 ± 2 mm Hg; P < 0.0001) and intra-abdominal LES length was shorter ( − 1.0 ± 0.3 cm, P < 0.0006) in GERD patients.

CONCLUSIONS: GERD patients have a wider esophagogastric insertion angle and have altered gastric morphology; structural changes that could compromise refl ux protection by the “ fl ap valve ” mechanism. In addi-tion, the EGJ opens wider during refl ux in GERD patients than in healthy volunteers: an effect that facilitates volume refl ux of gastric contents.

SUPPLEMENTARY MATERIAL is linked to the online version of the paper at http://www.nature.com/ajg

Am J Gastroenterol 2014; 109:658–667; doi: 10.1038/ajg.2014.25; published online 4 March 2014

1 Institute for Biomedical Engineering, University and ETH Zurich , Zurich , Switzerland ; 2 Division of Gastroenterology and Hepatology, University Hospital Zurich , Zurich , Switzerland ; 3 Department of Biological Sciences and Bioengineering, Indian Institute of Technology , Kanpur , India ; 4 Menne Biomed , Tuebingen , Germany ; 5 Division of Gastroenterology, Royal Melbourne Hospital , Melbourne , Australia ; 6 Zurich Centre for Integrative Human Physiology (ZIHP) , Zurich , Switzerland ; 7 Nottingham Digestive Diseases Centre and Biomedical Research Unit, University Hospital , Nottingham , UK ; � Deceased. Correspondence: Mark Fox, MD , Division of Gastroenterology and Hepatology, University Hospital Zurich , Raemistrasse 100 , CH-8091 Zurich , Switzerland . E-mail: [email protected] Initial observations from 12 GORD patients and 12 healthy subjects were presented at Digestive Diseases Week 2010. This work is dedicated to Dr Anupam Pal (deceased June 2012) who developed the 3D reconstruction of gastroesophageal anatomy from MR images and the novel biophysical analysis presented in this publication. Received 26 August 2013; accepted 27 January 2014

© 2014 by the American College of Gastroenterology The American Journal of GASTROENTEROLOGY

659

ES

OP

HA

GU

S

EGJ Structure and Function

protect against gastroesophageal refl ux (GER) despite a variety

of challenges, including large meals, abdominal straining, and

recumbent posture ( 1,2 ). Most refl ux events occur during spon-

taneous relaxations of the EGJ, called “ transient LES relaxations ”

(TLESRs), which are triggered by gastric distension, whose func-

tion in health is to vent air swallowed with the meal (belching)

( 3,4 ). A key observation in patients with mild-to-moderate gas-

troesophageal refl ux disease (GERD) is that TLESR frequency is

not greatly elevated compared with that in healthy controls ( 5,6 );

rather it is the likelihood of acid refl ux occurring during TLESR

that is a marker of disease ( 5 ). Further , manometry studies have

revealed no diff erences in the functional characteristics of TLESR

between healthy and GERD subjects (e.g., nadir pressure, dura-

tion of relaxation) ( 7,8 ). Th is indicates that other factors must

contribute to the occurrence of gastroesophageal refl ux.

Several authors have proposed that the “ functional anatomy ”

of the EGJ, in particular a “ fl ap valve ” at the entry of the esopha-

gus into the stomach, may provide additional protection against

refl ux ( 1,9,10 ); however, the data used to support this view were

acquired in anesthetized animals, in surgical preparations, or by

using invasive methodology that alter EGJ structure and function

( 11 – 14 ). Furthermore, these studies did not provide an adequate

description of the mechanism of refl ux protection: endoscopy

off ers a semi-qualitative assessment of EGJ morphology in the

fasted and gas-distended stomach, manometry delivers only an

indirect assessment of EGJ structure based on pressure (e.g., LES

length and position), and video fl uoroscopy suff ers from a lack of

soft -tissue contrast and exposes patients to ionizing radiation that

precludes prolonged or repeated measurements. Computed tom-

ography (CT) has been used to assess EGJ anatomy, including the

“ angle of His ” , with measurements obtained from 2D image slices

aft er ingestion of barium liquid ( 15 ). However, previous work has

shown that measurements of a complex 3D object such as the EGJ

on the basis of one 2D image do not provide reliable results because

any rotation of the object in the abdomen or selection of any other

slice plane would change the measurement outcome ( 16 ).

To overcome these diffi culties, we have developed and validated

concurrent magnetic resonance imaging (MRI) and high-resolu-

tion manometry (HRM) to obtain a comprehensive assessment

of the 3D “ functional anatomy ” of the EGJ in the fasted state and

aft er ingestion of a large test meal ( 16 ). Th ree-dimensional models

of the distal esophagus and proximal stomach were reconstructed

from stacks of 2D image slices from which objective measure-

ments were made including the insertion angle of the esophagus

into the stomach, gastric morphology, and orientation within the

abdomen ( 16,17 ). In addition, refl ux events were imaged in real

time by cine-magnetic resonance imaging (MRI) with concurrent

HRM providing a continuous assessment of EGJ function includ-

ing LES pressure and occurrence of TLESRs ( 16 ). Previously, based

on biophysical analysis of these data in healthy subjects, we pro-

posed a “ fl ap valve ” mechanism by which proximal gastric disten-

sion compresses the distal esophagus aft er a meal, a mechanism

that, if present, would increase resistance to gastroesophageal

refl ux and would protect the esophagus from refl ux even when the

EGJ is relaxed during TLESRs ( 17 ).

Th is study applies concurrent MRI and HRM to test the hypothesis

that structural factors thought to be involved in refl ux protection,

in particular the acute esophagogastric insertion angle, are altered

in patients with mild-to-moderate GERD.

METHODS Study population A total of 24 GERD patients (10 female, mean age 39 ± 14 years, mean

body mass index (BMI) 25.2 ± 3.5 kg / m 2 ) and 24 healthy volun-

teers (11 female, mean age 27 ± 5 years, mean BMI 22.6 ± 2.3 kg / m 2 )

participated in the study. GERD diagnosis was established by the

presence of refl ux esophagitis on endoscopy on the basis of LA

classifi cation and / or pathologic acid exposure on 24 h pH moni-

toring. Patients with a hiatus hernia ( > 3 cm), indicative of gross

EGJ disruption ( 18 ), were excluded as were those with a history

of other digestive diseases, abdominal surgery (except appendec-

tomy), or any condition that required active medical management.

Healthy subjects had no symptoms suggestive of refl ux, dyspepsia,

or other digestive problems and were not taking any medication

that could alter gastrointestinal function (endoscopy and pH studies

were not performed). Table 1 describes the study population. Th is

prospective clinical study was approved by the Ethics Committee

of Canton Zurich, Switzerland (EK 1631) and registered on www.

ClinicalTrials.gov (NCT01053585). Written informed consent was

obtained from all participants. All coauthors had access to the study

data and have reviewed and approved the fi nal manuscript.

Study design GERD patients discontinued PPI treatment at least 3 days prior

to the study. Alginate or antacid preparations were provided for

breakthrough symptoms. All participants attended the depart-

ment aft er a minimum 4 h of fasting. Th e structure and func-

tion of the proximal stomach and EGJ were assessed by validated

Table 1 . Demographic and clinical characteristics of the study population

Characteristic Healthy sub-jects ( n =24) GERD patients ( n =24)

Male / female 13 / 11 14 / 10

Age (range), years 27 ± 5 (20 – 35) 39 ± 14 (19 – 62)

BMI (kg / m 2 ) 22.6 ± 2.3 25.2 ± 3.5

GERD grade A (male / female) and / or pH > 6

— 14 (6 / 8)

GERD grade B — 3 (2 / 1)

GERD grade C — 7 (6 / 1)

Hiatus hernia ( < 3 cm) — 3 (3 / 0)

Duration of GERD symptoms a — 6 years (1 year – 20 years)

Duration of GERD diagnosis a — 1 year (1 month – 5 years)

Duration of GERD treatment a — 1 year (1 month – 5 years)

BMI, body mass index; GERD, gastroesophageal refl ux disease. a Before the inclusion in the study.


660 E

SO

PH

AG

US

Curcic et al.

methodology ( 16 ). A water-perfused, 21-channel silicone catheter

(Mui Scientifi c, Mississauga, ON, Canada) with an HRM system

custom-built for use with MRI (Royal Melbourne Hospital, Mel-

bourne, Australia) provided a continuous record of intra-luminal

pressure from the pharynx to the stomach. An MRI system, 1.5 T

Achieva (Philips Healthcare, Best, the Netherlands), with a four-

channel phased-array receive coil-obtained anatomic and cine-

images of the esophagus and stomach. Participants were studied

in the right decubitus position to reduce TLESR inhibition that

occurs on lying down ( 19 ). Aft er baseline HRM and MRI meas-

urements, a standardized mixed solid and liquid meal (600 g, 939

kcal, 44 % lipids) was ingested. Half an hour later, subjects drank a

liquid nutrient drink (0.5 kcal / ml) to maximum fullness (satiety).

Measurements of gastroesophageal anatomy and function were

obtained in four, 30-min study phases: fasting, aft er meal, fullness,

and emptying ( Figure 1 ).

Magnetic resonance imaging Details of the MRI acquisition protocol have been published ( 16 ).

To increase image contrast, all liquids were labeled with a para-

magnetic contrast agent (Dotarem, Guerbet, Villepinte, France).

Anatomic MRI consisted of 30 axial images positioned for high

spatial resolution reconstruction of the EGJ area and 30 sagit-

tal images covering the whole stomach for volume calculations.

Anatomic MRI was acquired during 15-s breath-holds in deep

inspiration and expiration. To visualize refl ux events, cine-MRI of

the distal esophagus was performed with a temporal resolution of

1 s simultaneously in three oblique coronal image slices orientated

along the esophageal axis over 6.8 min (limit of MRI workstation

memory).

High-resolution manometry HRM measurements were performed continuously during the

study. Th e HRM catheter (diameter 4 mm) had 21 pressure-

recording ports at 0.5 – 3 cm separation. A 6-cm segment of the

catheter assembly, with pressure-recording ports spaced 0.5 cm

apart, was positioned across the EGJ. Pressure data were recorded

at 25 samples per second.

Data analysis Magnetic resonance imaging . Image analysis was performed us-

ing algorithms developed in IDL 6.3. (ITT Visual Information

Solutions, Boulder, CO) and customized image analysis soft ware

developed in Matlab 7.11 (MathWorks, Natick, MA) as described

previously ( 16,17 ). Briefl y, the region of interest was segmented

on each anatomic MRI scan during breath-holds in inspiration

and expiration. From these contours, a 3D surface of the EGJ and

proximal stomach was generated with further analysis on the ba-

sis of these 3D models ( Figure 2 ).

Insertion angle . Th e HRM catheter indicated the esophageal axis.

By rotating a slice plane around this axis, 2D contours were gen-

erated at 1.8 ° intervals to provide a complete 360 ° description of

the “ insertion angle ” between the esophageal axis and a tangent

on the proximal fundus ( Figure 2a ) ( 16,17 ). Th e distribution of

the insertion angle shows a high frequency within 10 ° of the nadir

and the angle distribution under this 10 ° threshold is averaged to

HRM catheterplacement

Continuous HRM measurements

–15 Fasting After meal Fullness Emptying0 30 60 90 120

AnatomicMRI

Mixed meal

AnatomicMRI

AnatomicMRI

AnatomicMRI

Cine MRI

Liquid meal

Cine MRI Cine MRI

Time [min]

Figure 1 . Study protocol consisted of four phases: fasting, after meal, fullness, and emptying. Anatomic and cine-magentic resonance (MR) imaging was performed in each 30-min phase alternately.

Insertion angle Esophageal axis

T10 vertebrae

Proximalgastric

curvature

Sagittalmedian

lineEsophageal

insertionpoint

Major axisof the ellipsoid

T10 vertebrae

Stomachorientation

Figure 2 . Three-dimensional reconstruction of the esophagus and proxi-mal fundus (bright gray form) is superimposed on the 2D transverse ma-gentic resonance imaging (MRI) plane passing through the esophagogas-tric junction. ( a ) The red lines indicate the insertion angle defi ned as the angle between the esophageal axis and a tangent on the adjacent proximal stomach ( “ rotating plane ” analysis). Stomach orientation is the angle between the green line indicating the sagittal median line through the T10 vertebrae and the white line indicating the major axis of the proximal stomach ( “ ellipsoid fi t ” analysis). ( b ) Proximal gastric curvature is the arc over which the esophagus and the fundus are in close contact (within 10 ° of the insertion angle). The esophagogastric insertion point is the location at which the esophagus inserts into the proximal stomach, expressed as the angle between the major axis of the gastric ellipsoid and a line joining the center of the ellipsoid with the esophagus ( “ ellipsoid fi t ” analysis).


661

ES

OP

HA

GU

S


provide a stable measurement for statistical analysis. Th e rotating

plane analysis also defi nes the proximal gastric curvature (i.e. dis-

tension) as the arc over which the insertion angle remained below

the 10 ° threshold ( Figure 2b ).

Gastric morphology . Th e shape of the proximal fundus above the

EGJ was approximated by a least square fi t to an ellipsoid form

( 17 ). Th is analysis provided measurements of (i) the location at

which the esophagus inserts into the proximal stomach, expressed

as the angle between the major axis of the gastric ellipsoid and

a line joining the center of the ellipsoid with the center of the

esophagus (from here on called esophageal insertion point), and

(ii) the orientation of the proximal stomach within the abdo-

men, expressed as the angle between the major axis of the gastric

ellipsoid and the central abdominal axis passing through the center

of the T10 vertebrae ( Figure 2 ). Total gastric volume (i.e., gastric

content volume and intra-gastric air volume) was measured from

MR images to assess gastric fi lling and emptying.

Refl ux events . Cine-MRI was reviewed to detect refl ux events.

Th e duration of refl ux events was defi ned as the time from detec-

tion of labeled gastric content in the esophagus until its clearance

by esophageal contraction ( 16 ). EGJ opening during each refl ux

event was defi ned as the luminal diameter at the EGJ (the narrow-

est segment of the distal esophagus on imaging studies ( 20,21 ))

measured 3 s aft er retrograde fl ow commenced ( Figure 3 ).

High-resolution manometry . HRM pressure recordings were dis-

played by custom-made soft ware (Trace! 1.2, G.H.). LES pressure

and length were measured in each study phase from three separate,

10 s time windows that did not include swallows or TLESRs. LES

pressure was estimated at end expiration (5th centile of the 10-s

window) and in peak inspiration (95th centile) and referenced to

mean intra-gastric pressure. Th oracic and intra-abdominal LES

lengths were measured as the distance between the pressure in-

version point and the upper and lower LES borders, respectively.

TLESR and common cavity pressure events (equilibration of gas-

troesophageal pressure) were detected using criteria validated in

HRM systems ( 22 ). Th e rate of common cavity refl ux events dur-

ing the study was expressed as the number of events per hour.

Statistical analysis Th e primary outcome measure was the insertion angle of the

esophagus into the proximal stomach aft er the standard test

t = 0 s

Heart

Sto

mac

h

t = 1.013 s

t = 2.026 st = 2.026 s t = 3.039 s t = 4.052 s

Figure 3 . Six representative images of the beginning of one refl ux event as seen in the magnetic resonance imaging (MRI). Time point t = 0 is the moment when refl ux of gastric contents (bright, contrast agent-labeled fl uid) is fi rst seen in the image. Esophagogastric junction (EGJ) distension is measured ~ 3 s after t = 0. The insert demonstrates EGJ position, which is the narrowest segment of the distal esophagus, on MR images visualized during the refl ux event. This corresponded to the position of the lower esophageal sphincter on high-resolution manometry (note that the silicone catheter and pressure sensors on the catheter are clearly seen on images).


662 E

SO

PH

AG

US

Curcic et al.

duration, and EGJ opening, again with study phase as fi xed eff ects

and subjects as random eff ects. In this case, parameter estimates

are presented as means and 95 % confi dence intervals. Th e level of

signifi cance for all tests was 0.05.

RESULTS Study participants Demographic, clinical, and endoscopic characteristics of study

participants are presented in Table 1 . Healthy volunteers were

younger and had a slightly lower BMI compared with GERD

patients. No signifi cant eff ects of age or BMI were detected in

either study group on the esophagogastric insertion angle or sec-

ondary parameters ( Supplementary Figure 1 online).

Refl ux events and EGJ opening All refl ux events detected during cine-MRI were detected also on

HRM recordings ( Figure 4 ). Th e number of refl ux events aft er

the meal was similar in healthy volunteers and GERD patients

( Table 2 ). Healthy volunteers had on average of two refl ux events

aft er a meal (95 % CI: 1.8 – 3.3) and GERD patients had three refl ux

meal in expiration. Preliminary MRI studies recruited 12 GERD

patients and 12 healthy controls, and reported a wider (more

obtuse) insertion angle in the patient group ( + 7 ± 3 ° ; P < 0.017)

( 23 ). As these novel fi ndings needed confi rmation and we wished

to report additional biomarkers that provide insights into the

mechanism of refl ux protection ( 17 ), the fi nal study recruited 24

participants in each group. Th is study was designed and powered

to detect diff erences in EGJ structure and function between the

two defi ned study groups.

Statistical analysis was performed by linear mixed-eff ects mod-

eling using the soft ware package R 2.14.2 (R Foundation, Vienna,

Austria). Univariate linear mixed model analysis was used to evalu-

ate the eff ects of study group, respiration, and the four study phases

on the EGJ structural and functional parameters. Eff ects of study

group, respiration, and study phase were taken as fi xed eff ects

and subjects as random eff ect. Parameter e stimates and respec-

tive contrasts, i.e., the diff erences from the estimates of healthy

subjects, are presented with s.e. Th e number of refl ux events is a

count variable that follows a Poisson distribution. For this reason,

a univariate generalized linear model (log-linear model) was used

to evaluate diff erences in the number of refl ux events, refl ux event

UES

Reflux

Open LES

t = 283sLES

t = 270s

t = 270s

t = 279s

t = 279s

t = 292s

t = 283s t = 292s

–15–10–6–226101520253035404550

60

70

80

90

100

110

mmHg

Time

Dis

tanc

e fr

om n

ares

Peristalsis

Figure 4 . Refl ux events captured with both HRM (upper panel) and magnetic resonance imaging (MRI) (lower panel) in a gastroesophageal refl ux disease patient. A relatively short refl ux event was selected to facilitate presentation. Four time points corresponding to baseline prior to refl ux, commencement of refl ux, maximal esophageal fi lling, and clearance of the refl ux event captured with MRI are indicated by arrows on HRM recordings. Note that common cavity pressure was not uniform due to imperfect correction of radio-frequency interference of signal within MRI system.


663

ES

OP

HA

GU

S


events (95 % CI: 2.5 – 4.6) during a total observation time of ~ 20 min

( P = 0.092). Refl ux duration was longer in GERD patients by 6 s,

being on average 26.8 s (95 % CI: 23.9 – 30.2 s) in healthy subjects

and 32.8 s (95 % CI: 29.2 – 37.0 s) in GERD patients ( P = 0.002). EGJ

opening during refl ux was larger in GERD patients than in healthy

subjects. Th e EGJ diameter during refl ux averaged 16.8 mm (95 %

CI: 15.1 – 18.6 mm) in healthy subjects and 19.3 mm (95 % CI:

17.6 – 21.1 mm) in GERD patients ( P = 0.04).

Magnetic resonance imaging Measurements of EGJ structure and function are summarized in

Table 3 .

Esophagogastric insertion angle and proximal gastric disten-

sion . Th e insertion angle of the esophagus into the proximal

stomach was 55 ± 2.5 ° in the healthy group aft er meal intake and

during expiration. Th e insertion angle was wider (more obtuse) in

GERD patients by 7 ± 3 ° ( P = 0.03) ( Figure 5 , with animated im-

ages in Supplementary Videos 1 and 2 ). Gastric fi lling did not

alter this parameter. Respiration modulated the insertion angle in

both groups, being more obtuse (wider) during inspiration than

during expiration ( + 3 ± 1 ° ; P = 0.00001).

Proximal gastric curvature (i.e. distension) was 74 ± 4 ° in healthy

group and was similar in healthy controls and GERD patients

( + 2 ± 4 ° , P = 0.56). As expected, this measurement increased with

gastric fi lling from the fasted to the postprandial phase ( + 18 ± 4 ° ,

P < 0.0001) and increased further aft er feeding to maximum full-

ness ( + 8 ± 5 ° ). It did not change with respiration ( P = 0.32).

Gastric content volume and intra-gastric air volume . Gastric con-

tent volume directly aft er meal intake was similar in healthy controls

and GERD patients; however, intra-gastric air volume aft er the meal

was larger in healthy subjects than in the GERD group by 64 ± 29 ml

( P < 0.03). Gastric emptying during the 30-min emptying phase was

3.8 ± 0.4 ml / min in healthy volunteers, which was 1.6 ± 0.6 ml / min

faster compared with that in GERD patients ( P = 0.009).

Stomach orientation and esophagogastric insertion point . Th e

angle of orientation of the proximal stomach within the abdomen

relative to the central abdominal axis ( Figure 2 ) in the healthy

group aft er meal intake and during expiration was 60 ± 5 ° . Th is an-

gle was smaller in GERD patients by − 17 ± 7 ° ( P = 0.01), implying

that the proximal stomach is in a more anterior-posterior orienta-

tion in GERD ( Figure 6 ). Gastric orientation in the abdomen did

not change with gastric fi lling / study phase.

Th e insertion point of the esophagus relative to the major (long)

axis of the proximal stomach ( Figure 2 ) in the healthy group aft er

meal intake and during expiration was 31 ± 5 ° . Th e insertion point

was displaced further away from the major axis in GERD patients

by 12 ± 6 ° ( P = 0.05) ( Figure 6 ). Gastric fi lling had no eff ect on this

parameter.

Measurements of gastric orientation within the abdomen and

the position of insertion of the esophagus into the stomach are

strongly correlated during respiration, as the anatomical relation-

ship between the esophagus and proximal stomach is fi xed (Pear-

son correlation coeffi cient r = 0.8). Th us, during inspiration, as the

stomach rotated into a more antero-posterior position within the

Table 2 . Summary of observation times and of refl ux events detected

HRM MRI

Observation time [h]

Number of detected refl ux

events Median number of refl ux

events / h (interquartile range) Observation

time (h) Number of detected

refl ux events

HV 21 161 7.06 (3.48 – 9.98) 8.4 38

GERD 26 236 9.53 (5.56 – 15.0) 9.8 97

GERD, gastroesophageal refl ux disease; HRM, high-resolution manometry; HV, healthy volunteers; MRI, magnetic resonance imaging. HRM measurements were acquired continuously. Concurrent MRI measurements were acquired for about 40 % of the study duration. Agreement between the two modali-ties was perfect.

Table 3 . Biomarkers of esophageal and gastric structure and function.

Esophagogastric junction Gastric volumes Gastric morphology HRM measurement

Parameter

Esophagogastric insertion angle

Proximal gastric curvature / distension

Content (ml)

Air (ml)

Stomach orientation within abdomen

Esophago-gastric insertion point

LES pressure (mm Hg)

Intra-abdominal LES length (cm)

HV 55 74 723 165 60 31 20 1.9

GERD 62 76 662 101 43 43 8 0.9

Difference + 7 ( ± 3) 2 ( ± 4) − 61 ( ± 56) − 64 ( ± 29) − 17 ( ± 7) 12 ( ± 6) − 12 ( ± 2) − 1.0 ( ± 0.3)

P value 0.03 0.56 0.3 < 0.03 < 0.01 0.05 0.0001 0.0006

GERD, gastroesophageal refl ux disease; HRM, high-resolution manometry; HV, healthy volunteers; LES, lower esophageal sphincter. All measurements acquired in the “ after meal ” phase during expiration.


664 E

SO

PH

AG

US

Curcic et al.

DISCUSSION Th is study provides a novel and comprehensive description of

gastroesophageal structure in three dimensions by MRI with con-

current assessment of the EGJ function by HRM. Th e fi ndings

demonstrate that the “ functional anatomy ” of the refl ux barrier is

altered in patients with mild-moderate GERD.

Th e primary outcome of this study was that the insertion angle

of the esophagus into the stomach is wider (more obtuse) in GERD

patients than in healthy subjects aft er ingestion of a large test meal.

Th is anatomic relationship is referred to as the “ angle of His ” in

surgical studies ( 11 – 13 ), and may also be related to the tissue fold

on the lesser curve adjacent to the EGJ seen on endoscopy, which

is reduced in GERD patients ( 12,14 ). An acute esophagogastric

insertion angle is essential for the proposed “ fl ap valve ” mecha-

nism of refl ux protection, as this increases the mechanical advan-

tage by which the intra-abdominal esophagus is compressed by the

proximal stomach as it fi lls aft er a meal ( 17 ). Such compression

would limit EGJ opening during TLESRs and, as a result, reduce

or prevent refl ux. Consistent with the “ fl ap valve ” hypothesis, EGJ

opening during refl ux events was less in healthy volunteers than

in GERD patients (see below); however, it should be noted that

EGJ opening was variable between refl ux events and there was no

signifi cant association between the insertion angle and mean EGJ

opening. Th is lack of association could be because other factors are

more important to refl ux protection. Alternatively, the variability

of EGJ opening may indicate that active muscle contraction rather

than passive, structural factors contribute to the “ fl ap valve ” at the

EGJ. Specifi cally, active contraction of the sling and clasp fi bers

at the gastric cardia could maintain an acute insertion angle and

provide refl ux protection ( 2 ). Indeed, data from a recent pharma-

cologic intervention study demonstrate that refl ux suppression by

the GABA-B agonist baclofen is mediated not only by augment-

ing LES function but also by “ tightening ” (making more acute)

the insertion angle consistent with the “ fl ap valve ” mechanism of

refl ux protection ( 24 ).

Proximal gastric curvature, a parameter that refl ects distension

of the stomach aft er a meal, increased in all subjects as the stomach

fi lled from an elliptical (empty) to a more spherical (full) form.

Additional MRI measurements identifi ed the point at which the

esophagus inserts into the stomach, the position of the stomach

within the abdomen, and the dynamic change in gastric volume

and morphology aft er a meal. In healthy subjects, the stomach is

orientated laterally within the abdomen and the esophagus enters

the stomach high on the lesser curve, close to the major axis of the

gastric ellipse. Gastric morphology in GERD patients is heteroge-

neous ( Supplementary Figure 2 ); however, as a group, the stom-

ach was found to occupy an abnormal antero-posterior position in

the abdomen with the esophageal insertion point displaced later-

ally from the major axis of the ellipse ( Figure 6 ). Th e biophysical

model described by Roy et al. ( 17 ) predicts that, in healthy subjects

with normal gastric morphology, the increase in proximal disten-

sion that occurs as the stomach fi lls during the meal promotes

compression of the distal esophagus and augments the proposed

“ fl ap valve ” . In contrast, in GERD patients with an eccentric inser-

tion point of the esophagus, contact between the proximal stomach

abdomen by 6.0 ± 1.6 ° ( P = 0.0004), the insertion point appeared to

decrease by − 7 ± 2 ° ( P < 0.003). Further, within the study popula-

tion, there was a relationship between stomach orientation and

the insertion point ( Supplementary Figure 2 ).

High-resolution manometry LES / EGJ pressure . LES pressure aft er meal intake and during

expiration in healthy volunteers was 20 ± 2 mm Hg. It was lower

in GERD patients by 12 ± 2 mm Hg ( P < 0.0001). EGJ pressure

decreased aft er meal ingestion in both groups ( P = 0.05), and re-

mained lower than baseline during fullness and emptying phases.

LES length . Intra-abdominal LES length aft er meal intake was

1.9 ± 0.2 cm in healthy subjects and was shorter in GERD patients

by 1.0 ± 0.3 cm ( P < 0.0006). Th oracic LES length was not diff er-

ent between the groups ( P = 0.51). No signifi cant change in LES

length was detected during the study phases.

Meanesophago-gastric

insertion anglein health

Meanesophago-gastric

insertion anglein GERDs.d.

s.d.~55°

Δ = +7°

Fundus Fundus

Antrum

Heart Heart

Proximalfundus

StomachStomach

Liver

Catheter

Eso

phag

us

Eso

phag

usAntrum

Figure 5 . Visualization of the insertion angle. ( a ) Schematic representa-tion of the mean esophagogastric insertion angle in healthy volunteers and gastroesophageal refl ux disease (GERD) patients . In healthy volunteers, the dark blue line is the esophageal axis, whereas the black lines form the tangent on the fundus. Dashed blue lines are standard deviations of the group. The insertion angle in GERD patients (light red line) is overlaid for comparison. In GERD patients, mean esophagogastric insertion angle between the esophageal axis (dark red) and the proximal stomach (black) in GERD patients is more obtuse, by 7 ° , with respect to that in healthy con-trols (light blue line overlaid for comparison). ( b ) Representative examples of an acute insertion angle in healthy subjects (left), ~ 25 ° , and an obtuse angle in GERD (right), ~ 70 ° . The vertical plane passes through the esoph-agogastric junction (EGJ) at the level of the T10 vertebra and the horizontal plane is orientated to pass through the long axis of the stomach. Note that the vertical plane is rotated into a more antero-posterior position because of the abnormal position of the stomach in the abdomen in this GERD patient. See also animated images in Supplementary Videos 1 and 2 .


665

ES

OP

HA

GU

S


and the intra-abdominal esophagus does not increase during fi ll-

ing and this undermines the protection provided by the fl ap valve

mechanism. In this situation, failure of refl ux protection by the fl ap

valve mechanism occurs because the relative change in curvature

and, therefore, compression of the esophagus by the stomach is

reduced as the esophagogastric insertion point moves away from

the major axis of the stomach ( 17 ).

Direct observations by HRM and concurrent cine-MRI had

perfect concordance for refl ux detection ( 16 ). Most refl ux events

occurred during TLESRs. As expected ( 5,6 ), there was little diff er-

ence in the frequency of TLESRs and refl ux events in healthy indi-

viduals and GERD patients. However, important diff erences were

still present between the study groups. First, EGJ opening during

refl ux events was signifi cantly less in healthy volunteers than it was

in GERD patients by an average of 0.25 cm. Th is is consistent with

the fi nding of increased EGJ compliance in GERD reported by

barostat studies ( 18 ); however, this is the fi rst time that increased

EGJ opening in GERD patients has been documented in vivo dur-

ing refl ux events by noninvasive imaging. According to Hagen-

Poiseuille ’ s law, this seemingly small (approximately + 15 % )

diff erence in luminal diameter would result in a relatively large

increase in refl ux volume (approximately + 75 % ) because fl ow is

proportional to the radius to the fourth power. Th is is much greater

than the 1:1 increase in fl ow that would be related to reduced LES

nadir pressure, LES length, or increased abdominal pressure. Sec-

ond, once refl ux had occurred, esophageal volume clearance was

signifi cantly slower in GERD patients. Together, these observa-

tions provide a possible, mechanical explanation for the increase

in “ volume refl ux ” and for prolonged refl ux exposure that cause

symptoms and mucosal damage in this condition ( 25,26 ).

Postprandial stomach function was assessed by the change in

gastric volume aft er ingestion of the test meal. Th e initial increase

in gastric volume was similar in both groups; however, subsequent

emptying was ~ 40 % slower in GERD patients than in healthy

volunteers. Th ese fi ndings are consistent with selected reports

from barostat studies of slow recovery of gastric tone aft er a meal

( 27,28 ). Th is slow gastric emptying could lead to retention of the

meal and secretions in the proximal stomach, and thus increase the

likelihood of refl ux. In addition, delayed gastric emptying would

prolong the time over which refl ux can occur ( 29,30 ).

Th is study provides objective evidence that gastroesophageal

anatomy as well as function is abnormal in GERD patients and

may also provide insights into the mechanism of refl ux protec-

tion. It is clear that the high-pressure zone at the EGJ prevents

refl ux when the EGJ barrier is closed; however, low resting EGJ

pressure does not explain why GERD patients are more likely than

healthy controls to have refl ux during TLESRs ( 7,8 ). In contrast,

the proposed “ fl ap valve ” formed by an acute esophagogastric

insertion angle represents a mechanism of refl ux protection that

could operate during TLESRs aft er meals. As described above, an

acute insertion angle would allow the distended proximal stomach

to compress the intra-abdominal esophagus, limit EGJ opening,

and increase resistance to retrograde fl ow (i.e. refl ux) even when

the LES is relaxed. Larger studies are required to assess whether

an acute insertion angle or other factors are the cause of reduced

refl ux; however, irrespective of the mechanism, the fi nding that

EGJ opening during TLESRs is greater in GERD patients than in

healthy controls is likely to be clinically relevant. Th is is because

wider EGJ opening during retrograde fl ow would allow larger vol-

ume and more proximal refl ux events that are more likely to cause

symptoms, in particular regurgitation ( 26 ), which are not well con-

trolled by proton pump inhibitors ( 31 ).

Th is study applied MRI to assess not only esophagogastric struc-

ture but also function. In contrast to video fl uoroscopy or CT

imaging, MRI provides excellent soft tissue contrast and allows

repeated measurements before and aft er eating a meal as well as

cine-MRI of refl ux events without the risks of ionizing radiation.

Th is high-resolution imaging, combined with 3D reconstruction

Ellipsoid fit of thefundus above EGJ:

healthy subjects

Esophagealinsertion point

Esophagealinsertion point

T10 vertebrae

Stomachorientation

Stomchorientation

P P

STD STD

STDSTD

O O

Minor

axis

Major axis

Esophagus Esophagus

A

Ellipsoid fit of thefundus above EGJ:

GERD patients

Minor axis

Major axis

A

Figure 6 . Visualization of the stomach orientation. Caudal view of the mean ellipsoid fi t of proximal stomach at the level of the esophagogastric junction (EGJ) after the meal and during expiration in healthy subjects ( a ) and in gastroesophageal refl ux disease (GERD) patients ( b ). The insertion point of the es-ophagus into the stomach and the orientation of the proximal stomach in the abdomen are indicated. Dashed dark gray lines are the s.d. Dashed light gray lines are the major and minor axes of the ellipsoid fi t. In healthy controls, the esophagus enters the proximal stomach close to the major axis, whereas in GERD patients it is displaced away from the major axis. It follows that, as the esophagus is fi xed in the diaphragm hiatus, the proximal stomach is oriented in a more anterio-posterior position within the abdomen in GERD patients than it is in healthy subjects.


666 E

SO

PH

AG

US

Curcic et al.

S. Hebbard: technical support; critical revision of the manuscript

for important intellectual content; Reto Treier: MRI-protocol setup;

Peter Boesiger: obtained funding, technical or material support of

MRI, and fi nal manuscript approval; Andreas Steingoetter: MRI

support; technical supervision; Michael Fried: obtained funding,

and fi nal manuscript approval; Werner Schwizer: obtained funding,

study supervision, and fi nal manuscript approval.

Financial support: Th is investigator-led research was supported by

an unrestricted research grant from Astra Zeneca (Molndal, Sweden).

Th e authors retained control of the data and the information

submitted for publication. Additional fi nancial support was provided

by the Baugarten Foundation of the University of Zurich and the

Zurich Centre for Integrative Human Physiology. A.P. was supported

by initiation grant from Indian Institute of Technology Kanpur,

and Council of Scientifi c and Industrial Research, India (Scheme

number: 37(1287) / 07 / EMR-II). Th e collaboration between M.F. and

A.P. is supported by a Royal Society International Exchange Grant

( # 2011 / 0729032650226750).

Potential competing interests: None.

Study Highlights

WHAT IS CURRENT KNOWLEDGE 3 The esophagogastric junction (EGJ) is a key defense against

refl ux.

3 Most refl ux in healthy subjects and in patients with mild-moderate gastroesophageal refl ux disease (GERD) occurs during transient lower esophageal sphincter relaxations (TLESRs).

3 However, TLESR frequency is not necessarily increased in GERD, rather it is the likelihood of refl ux occurring during TLESRs that is increased.

3 Structural factors including a “ fl ap valve ” at the EGJ may protect against refl ux during TLESRs; however, this is diffi cult to assess with the current, invasive methodology that alters EGJ structure and function.

WHAT IS NEW HERE 3 Minimally invasive technology that combines magnetic

resonance imaging (MRI) and high-resolution manometry (HRM), and provides a comprehensive assessment of esophagogastric structure and function after a large test meal.

3 Compared with healthy subjects, GERD patients have a wider esophagogastric insertion angle and other abnormali-ties of gastric anatomy that can compromise refl ux protec-tion by the “ fl ap valve ” mechanism.

3 GERD patients also have wider EGJ opening during refl ux events after meals, allowing greater volume of refl ux events.

3 Abnormal gastric anatomy on magnetic resonance imag-ing could identify GERD patients with pathological acid exposure.

REFERENCES 1 . Mittal RK , Balaban DH . Th e esophagogastric junction . N Engl J Med

1997 ; 336 : 924 – 32 . 2 . Brasseur JG , Ulerich R , Dai Q et al. Pharmacological dissection of the

human gastro-oesophageal segment into three sphincteric components . J Physiol 2007 ; 580 : 961 – 75 .

of esophagogastric anatomy and validated image analysis tech-

niques ( 17 ) provides objective measurements, which are superior

to the selective reporting of the 2D images previously reported

( 15 ). Th is novel methodology was applied in a group of GERD

patients without large hiatus hernia. Th e results were consistent

with the study hypothesis; however, it was neither designed nor

powered to confi rm a causal association between measurements

of EGJ structure / function and clinical measurements (e.g. number

of refl ux events and esophageal acid exposure). Further, it should

be noted that the patient group was 12 years older and somewhat

heavier (BMI + 2.5 kg / m 2 ) compared with the controls. No signifi -

cant correlation was present between these variables and the meas-

urements of EGJ structure ( Supplementary Figure 1 ). Further,

published data from large patient cohorts suggest that age- and

weight-related changes are not suffi cient to explain the diff erence

in manometric fi ndings ( 15,32,33 ); however, this possibility can-

not be excluded. Th e main limitation of the MRI methodology

is that subjects are investigated lying down and not in the physi-

ologic, upright position. Inhibition of TLESRs is less in the right

than in the left decubitus position ( 19 ), but eff ects on the intra-

gastric distribution of meal and secretion cannot be avoided. As

a result, the EGJ was always in contact with the large test meal,

and therefore all TLESRs were associated with refl ux in the supine

position. Position may also aff ect the orientation of the stomach

within the abdomen; however, as demonstrated for respiration,

this would not alter the anatomical relation between the esopha-

gus and the stomach. Finally, current MRI protocols do not resolve

the gastric muscle layers, including the clasp and sling fi bers of the

cardia, which must be quantifi ed for a full biophysical description

of refl ux protection ( 2 ). Ongoing studies, will establish whether

the biomarkers described in this paper that characterize the func-

tional anatomy of the stomach discriminate GERD patients from

those with functional dyspepsia and other causes of postprandial

symptoms.

In conclusion, high-resolution MR imaging with innovative 3D

image analysis provided comprehensive information about gastro-

esophageal structure and function in health and disease. GERD

patients were shown to have a wider (more obtuse) esophagogas-

tric insertion angle and altered gastric morphology. Further, direct

observations by cine-MRI showed wider opening of the EGJ dur-

ing refl ux events in GERD patients, which would reduce resistance

to retrograde fl ow and increase the volume of refl ux events, com-

pared with healthy volunteers.

CONFLICT OF INTEREST Guarantor of the article: Mark Fox, MD.

Specifi c author contributions: Mark Fox: study concept and design,

obtained funding, study supervision, and draft ing of the manuscript;

Jelena Curcic: MRI-protocol setup, acquisition, analysis, and

interpretation of MRI-data, statistical analysis, and draft ing of the

manuscript; Shammodip Roy: analysis and interpretation of MRI-

data; Alexandra Schwizer: acquisition, analysis, and interpretation of

data; Elad Kaufman: acquisition, analysis, and interpretation of data;

Zsofi a Forras-Kaufman: acquisition, analysis, and interpretation of

data; Dieter Menne: statistical and scientifi c supervision; Geoff rey


667

ES

OP

HA

GU

S


3 . Holloway RH , Hongo M , Berger K et al. Gastric distention: a mechanism for postprandial gastroesophageal refl ux . Gastroenterology 1985 ; 89 : 779 – 84 .

4 . Dent J , Holloway RH , Toouli J et al. Mechanisms of lower oesophageal sphincter incompetence in patients with symptomatic gastrooesophageal refl ux . Gut 1988 ; 29 : 1020 – 8 .

5 . Trudgill NJ , Riley SA . Transient lower esophageal sphincter relaxations are no more frequent in patients with gastroesophageal refl ux disease than in asymptomatic volunteers . Am J Gastroenterol 2001 ; 96 : 2569 – 74 .

6 . Grossi L , Ciccaglione AF , Travaglini N et al. Transient lower esophageal sphincter relaxations and gastroesophageal refl ux episodes in healthy sub-jects and GERD patients during 24 hours . Dig Dis Sci 2001 ; 46 : 815 – 21 .

7 . Bredenoord AJ , Weusten BLAM , Carmagnola S et al. Double-peaked high-pressure zone at the esophagogastric junction in controls and in patients with a hiatal hernia: a study using high-resolution manometry . Dig Dis Sci 2004 ; 49 : 1128 – 35 .

8 . Scheff er RCH , Gooszen HG , Hebbard GS et al. Th e role of transsphinc-teric pressure and proximal gastric volume in acid refl ux before and aft er fundoplication . Gastroenterology 2005 ; 129 : 1900 – 9 .

9 . Atkinson M . Mechanisms protecting against gastro-oesophageal refl ux: a review . Gut 1962 ; 3 : 1 – 15 .

10 . McMahon BP , Jobe BA , Pandolfi no JE et al. Do we really understand the role of the oesophagogastric junction in disease? World J Gastroenterol 2009 ; 15 : 144 – 50 .

11 . Bardaji C , Boix-Ochoa J . Contribution of the His angle to the gastro-es-ophageal antirefl ux mechanism . Pediatr Surg Int 1986 ; 1 : 172 – 6 .

12 . Hill LD , Kozarek RA , Kraemer SJ et al. Th e gastroesophageal fl ap valve: in vitro and in vivo observations . Gastrointest Endosc 1996 ; 44 : 541 – 7 .

13 . Fujiwara Y , Nakagawa K , Kusunoki M et al. Gastroesophageal refl ux aft er distal gastrectomy: possible signifi cance of the angle of His . Am J Gastro-enterol 1998 ; 93 : 11 – 5 .

14 . Pandolfi no JE , Bianchi LK , Lee TJ et al. Esophagogastric junction morphol-ogy predicts susceptibility to exercise-induced refl ux . Am J Gastroenterol 2004 ; 99 : 1430 – 6 .

15 . Fukazawa K , Furuta K , Adachi K et al. Continuous imaging of esophago-gastric junction in patients with refl ux esophagitis using 320-row area detector CT: a feasibility study . J Gastroenterol Hepatol 2013 ; 28 : 1600 – 7 .

16 . Curcic J , Fox M , Kaufman E et al. Gastroesophageal junction: structure and function as assessed by using MR imaging . Radiology 2010 ; 257 : 115 – 24 .

17 . Roy S , Fox MR , Curcic J et al. Th e gastro-esophageal refl ux barrier: biophysical analysis on 3D models of anatomy from magnetic resonance imaging . Neurogastroenterol Motil 2012 ; 24 : 616 – 25 .

18 . Pandolfi no JE , Shi G , Trueworthy B et al. Esophagogastric junction opening during relaxation distinguishes nonhernia refl ux patients, hernia patients, and normal subjects . Gastroenterology 2003 ; 125 : 1018 – 24 .

19 . van Herwaarden MA , Katzka DA , Smout AJ et al. Eff ect of diff erent recum-bent positions on postprandial gastroesophageal refl ux in normal subjects . Am J Gastroenterol 2000 ; 95 : 2731 – 6 .

20 . Cohen M , Ellett MC , Perkins S et al. Accurate localization of the position of the tip of a naso/orogastric tube in children; where is the location of the gastro-esophageal junction? Pediatric Radiol 2011 ; 41 : 1266 – 71 .

21 . Cohen BR , Wolf BS , Som M et al. Correlation of Manometric, Oesophago-scopic, and Radiological Findings in the Columnar-Lined Gullet (Barrett Syndrome) . Gut 1963 ; 4 : 406 – 12 .

22 . Roman S , Zerbib F , Belhocine K et al. High resolution manometry to detect transient lower oesophageal sphincter relaxations: diagnostic accuracy compared with perfused-sleeve manometry, and the defi nition of new detection criteria . Aliment Pharmacol Th er 2011 ; 34 : 384 – 93 .

23 . Kaufman E , Curcic J , Pal A et al. Th e structure and function of the gastro-esophageal junction in health and refl ux disease assessed by magnetic resonance imaging and high resolution manometry . Gastroenterology 2009 ; 136 (Suppl 1) : A-17 .

24 . Curcic J , Schwizer A , Roy S et al. Eff ect of baclofen on the mechanism of refl ux protection in healthy volunteers and patients with gastro-esophageal refl ux disease (GERD) assessed by magnetic resonance imaging (MRI) and high resolution manometry (HRM): a randomized controlled study . Gastroenterology 2013 ; 142 : S – 304 .

25 . Lundell LR , Dent J , Bennett JR et al. Endoscopic assessment of oesophagi-tis: clinical and functional correlates and further validation of the Los Angeles classifi cation . Gut 1999 ; 45 : 172 – 80 .

26 . Bredenoord AJ , Weusten BL , Curvers WL et al. Determinants of perception of heartburn and regurgitation . Gut 2006 ; 55 : 313 – 8 .

27 . Penagini R , Hebbard G , Horowitz M et al. Motor function of the proximal stomach and visceral perception in gastro-oesophageal refl ux disease . Gut 1998 ; 42 : 251 – 7 .

28 . Zerbib F , des Varannes SB , Ropert A et al. Proximal gastric tone in gastro-oesophageal refl ux disease . Eur J Gastroenterol Hepatol 1999 ; 11 : 511 – 5 .

29 . McCallum RW , Berkowitz DM , Lerner E . Gastric emptying in patients with gastroesophageal refl ux . Gastroenterology 1981 ; 80 : 285 – 91 .

30 . Schwizer W , Hinder RA , DeMeester TR . Does delayed gastric emptying contribute to gastroesophageal refl ux disease? Am J Surg 1989 ; 157 : 74 – 81 .

31 . Kahrilas PJ , Jonsson A , Denison H et al. Regurgitation is less responsive to acid suppression than heartburn in patients with gastroesophageal refl ux disease . Clin Gastroenterol Hepatol 2012 ; 10 : 612 – 9 .

32 . Lee J , Anggiansah A , Anggiansah R et al. Eff ects of age on the gastro-esophageal junction, esophageal motility, and refl ux disease . Clin Gastroen-terol Hepatol 2007 ; 5 : 1392 – 8 .

33 . Anggiansah R , Sweis R , Anggiansah A et al. Th e eff ects of obesity on oesophageal function, acid exposure and the symptoms of gastro-oesopha-geal refl ux disease . Aliment Pharmacol Th er 2013 ; 37 : 555 – 63 .

Abnormal Structure and Function of the Esophagogastric Junction and Proximal Stomach in...

Documents

Transcript of Abnormal Structure and Function of the Esophagogastric Junction and Proximal Stomach in...